Microstructural engineering of modern high strength low alloy ...

UNSW Sydney

Microstructural engineering of modern high strength low alloy steels via

advanced thermo-mechanical processing

Carina Ledermueller

School of Materials Science and Engineering

Faculty of Science

September 2019

A thesis in fulfilment of the requirements for the degree of

Doctor of Philosophy

Thesis/Dissertation Sheet

Surname/Family Name : Ledermueller

Given Name/s : Carina

Abbreviation for degree as give in the University calendar : PhD

Faculty : Science

School : Materials Science & Engineering

Thesis Title : Microstructural engineering of modern high strength low alloy steels via advanced thermo-mechanical processing

Engineering microstructures in high strength low alloy (HSLA) steels via advanced thermo-mechanical processing (aTMP) is a

promising approach to overcome challenges around low work hardenability and toughness in ultrafine-grained mild steels. Thus,

the applicability of advanced thermo-mechanical processing for achieving multi-scale hierarchical microstructures in a HSLA steel

was studied. The microstructural evolution during warm deformation of a martensitic/bainitic starting microstructure using a

Gleeble 3500 thermo-mechanical simulator at 600°C followed by a direct aging step was investigated. A strain rate of 10 s-1

during a single pass plane strain compression led to strain localisation and, therefore, the formation of a macroscopic shear band.

Hence, an optimised advanced multi-hit thermo-mechanical process for achieving homogenous hierarchical microstructures

without strain localisation was developed. Ultrafine crystallites confined by a mixture of high angle gain and subgrain boundaries

are formed, decorated by two types of precipitates. Large FeMnC-rich cementite particles are found on grain boundaries and

smaller TiNbC-rich precipitates on dislocations and subgrain boundaries. The further aim of this study was the investigation of

the mechanical properties by upscaling the process developed in the Gleeble using a Hille 100 rolling mill. It was found that rolling

to reduction of thickness of 55% at a temperature around 650°C can lead to an ultimate tensile strength (UTS) of 650 MPa, a

yield ratio of 0.95 and a total elongation of 14% in the as-rolled condition. Delaminations did occur in the lower temperature region

of the Charpy impact testing in longitudinal and transversal directions. Finally, a similar processing was implemented to invoke

grain refinement coupled with strengthening arising from microalloying. It was found that Mo leads to an increase in hardness of

~20% compared to the base alloy, whereas Cr provides only a minor hardening increment. It was found that Mo is more effective

than Cr in delaying dislocation recovery. It was also observed that Mo partitions into nanoscale Nb-C solute clusters and

precipitates of NbC and Fe3C during ageing, retarding the coarsening of these phases. However, Cr was found to partition into

Fe3C only, and does not contribute to the nature of the dispersion of clusters and NbC.

Declaration relating to disposition of project thesis/dissertation I hereby grant to the University of New South Wales or its agents the right to archive and to make available my thesis or dissertation in whole or in part in the University libraries in all forms of media, now or here after known, subject to the provisions of the Copyright Act 1968. I retain all property rights, such as patent rights. I also retain the right to use in future works (such as articles or books) all or part of this thesis or dissertation. I also authorise University Microfilms to use the 350 word abstract of my thesis in Dissertation Abstracts International (this is applicable to doctoral theses only). …………………………………………………………… Signature

……………………………………..……………… Witness Signature

……….……………………...…….… Date

The University recognises that there may be exceptional circumstances requiring restrictions on copying or conditions on use. Requests for restriction for a period of up to 2 years must be made in writing. Requests for a longer period of restriction may be considered in exceptional circumstances and require the approval of the Dean of Graduate Research.

FOR OFFICE USE ONLY Date of completion of requirements for Award:

i

INCLUSION OF PUBLICATIONS STATEMENT

UNSW is supportive of candidates publishing their research results during their candidature

as detailed in the UNSW Thesis Examination Procedure.

Publications can be used in their thesis in lieu of a Chapter if:

The student contributed greater than 50% of the content in the publication and is the

“primary author”, ie. the student was responsible primarily for the planning, execution and

preparation of the work for publication

The student has approval to include the publication in their thesis in lieu of a Chapter from

their supervisor and Postgraduate Coordinator.

The publication is not subject to any obligations or contractual agreements with a third

party that would constrain its inclusion in the thesis

Please indicate whether this thesis contains published material or not.

☐ This thesis contains no publications, either published or submitted for publication (if this box is checked, you may delete all the material on page 2)

☐

Some of the work described in this thesis has been published and it has been documented in the relevant Chapters with acknowledgement (if this box is checked, you may delete all the material on page 2)

☒

This thesis has publications (either published or submitted for publication) incorporated into it in lieu of a chapter and the details are presented below

CANDIDATE’S DECLARATION

I declare that:

I have complied with the Thesis Examination Procedure

where I have used a publication in lieu of a Chapter, the listed publication(s) below meet(s) the requirements to be included in the thesis.

Name

Carina Ledermueller

Signature Date (dd/mm/yy)

20/08/20

Postgraduate Coordinator’s Declaration (to be filled in where publications are used in lieu of Chapters)

I declare that:

the information below is accurate

where listed publication(s) have been used in lieu of Chapter(s), their use complies with the Thesis Examination Procedure

the minimum requirements for the format of the thesis have been met.

PGC’s Name

Nagarajan Valanoor

PGC’s Signature Date (dd/mm/yy)

20/08/20

ii

For each publication incorporated into the thesis in lieu of a Chapter, provide all of

the requested details and signatures required

Details of publication #1: Full title: Engineering Hierarchical Microstructures via Advanced Thermo-mechanical Processing of a Modern HSLA steel Authors: Carina Ledermueller, Huijun Li, Sophie Primig Journal or book name: Metallurgical and Materials Transactions A Volume/page numbers: Volume 49, Issue 12, pp.6337-6350 Date accepted/ published: published online 11/09/2018

Status Published X Accepted and In press

In progress (submitted)

The Candidate’s Contribution to the Work C. Ledermueller did the majority of the planning and execution of the experimental work as well as the data analysis. The manuscript was drafted by her.

Location of the work in the thesis and/or how the work is incorporated in the thesis: Experimental work is described in Chapter 3, results and discussion in Chapter 4, parts of the abstract/introduction used in the thesis abstract and Chapter 2 & 8.

Primary Supervisor’s Declaration I declare that: • the information above is accurate • this has been discussed with the PGC and it is agreed that this publication can be

included in this thesis in lieu of a Chapter • All of the co-authors of the publication have reviewed the above information and have

agreed to its veracity by signing a ‘Co-Author Authorisation’ form.

Supervisor’s name Sophie Primig

Supervisor’s signature Date (dd/mm/yy) 20/08/20

Details of publication #2: Full title: Advanced thermo-mechanical process for homogenous hierarchical microstructures in HSLA steels Authors: C. Ledermueller, E. Kozeschnik, R. Webster & S. Primig Journal or book name: Metallurgical and Materials Transactions A Volume/page numbers: Volume 50, Issue 12, pp 5800-5815 Date accepted/ published: published online 11/10/2019

Status Published X Accepted and In press



Location of the work in the thesis and/or how the work is incorporated in the thesis: Parts of the experimental work is described in Chapter 3, results and discussion in Chapter 5, parts of the abstract/introduction used in the thesis abstract and Chapter 1, 2 & 8.






iii

Details of publication #3: Full title: An Initial Report on the Structure–Property Relationships of a High-Strength Low-Alloy Steel Subjected to Advanced Thermomechanical Processing in Ferrite Authors: C. Ledermueller, H. Zhu, H. Li, S. Primig Journal or book name: Steel Research International Volume/page numbers: Volume 91, Issue 7, Article Number: 1900596 Date accepted/ published: published online 05/02/2020

Status Published x Accepted and In press



Location of the work in the thesis and/or how the work is incorporated in the thesis: Parts of the experimental work described in Chapter 3, results and discussion in Chapter 6, parts of the abstract/introduction used in the thesis abstract and Chapter 2 & 8.






Details of publication #4: Full title: Microalloying effects of Mo versus Cr in HSLA steels with ultrafine grained ferrite microstructures Authors: C. Ledermueller, H. I. Pratiwi, R. F. Webster, M. Eizadjou, S. P. Ringer & S. Primig Journal or book name: Materials & Design Volume/page numbers: Volume 185, Article Number 108278 Date accepted/ published: published online 18/10/2019

Status Published x Accepted and In press


The Candidate’s Contribution to the Work C. Ledermueller did a major part of the experimental work as well as the data analysis. The manuscript was drafted by her. She supervised the work done by H.I. Pratiwi.

Location of the work in the thesis and/or how the work is incorporated in the thesis: Parts of the experimental work described in Chapter 3, results and discussion in Chapter 7, parts of the abstract/introduction used in the thesis abstract and Chapter 2 & 8.






i

ORIGINALITY STATEMENT

‘I hereby declare that this submission is my own work and to the best of my knowledge

contains no materials previously published or written by another person, or substantial

proportions of material which have been accepted for the award of any other degree or

diploma at UNSW or any other educational institution, except where due

acknowledgement is made in the thesis. Any contribution made to the research by others,

with whom I have worked at UNSW or elsewhere, is explicitly acknowledged in the thesis.

I also declare that the intellectual content of this thesis is the product of my own work,

except to the extent that assistance from others in the project’s design and conception or

in style, presentation and linguistic expression is acknowledged.’

Signed …………………………………………………….

Date ………………………………………………….

ii

COPYRIGHT STATEMENT ‘I hereby grant the University of New South Wales or its agents the right to archive

and to make available my thesis or dissertation in whole or part in the University

libraries in all forms of media, now or here after known, subject to the provisions

of Copyright Act 1968. I retain all property rights, such as patent rights. I also retain

the right to use in future works (such as articles or books) all or part of this thesis

or dissertation.

I also authorise University Microfilms to use the 350 word abstract of my thesis in

Dissertation Abstract International (this is applicable to doctoral theses only).

I have either used no substantial proportions of copyright material in my thesis or

I have obtained permission to use copyright material; where permission has not

been granted I have applied/will apply for a partial restriction of the digital copy of

my thesis or dissertation.’

Signed ………………………………………..

Date …………………………………………..

AUTHENTICITY STATEMENT

‘I certify that the Library deposit digital copy is a direct equivalent of the final

officially approved version of my thesis. No emendation of content has occurred

and if there are any minor variations in formatting, they are result of the conversion

to digital format.’

Signed ………………………………………..

Date …………………………………………..

iii

Acknowledgements

Undertaking a PhD was never part of my career planning. However, when the

opportunity arose to study overseas under the supervision of my former supervisor

Dr Sophie Primig I decided to dare to take on the adventure moving to the other

side of the world. Although PhD life was not always easy I am grateful for making

this decision. It was truly life-changing as it has opened up my mind to all that

beautiful variety of cultures and people out there. That said I would like to thank

everybody who was part of my journey during the last 3.5 years and has helped

me to complete my thesis.

Firstly, I would like to express my sincere gratitude to my supervisor, A/Prof Sophie

Primig, for her support, patience, motivation and immense knowledge. I am

grateful to be able to call her my supervisor and mentor all these years since

starting my undergraduate degree.

My sincere appreciation also goes to all the technical and professional staff at

UNSW Sydney and the Electron Microscopy Unit at UNSW but especially to

Prof Paul Munroe, Prof Michael Ferry, Dr Simon Hager, Dr Charlie Kong,

Dr Qiang Zhu, Dr George Yang and Dr Rahmat Kartono. I would like to

acknowledge Bill Joe’s help with the mechanical testing, Dr David Miskovic’s help

with the Gleeble experiments and Dr Richard Webster for the TEM imaging.

I further want to thank my collaborators at the University of Wollongong who have

enabled me to use their facilities. Prof Huijun Li and Dr Liang Chen for their support

and help with some of the Gleeble experiments and Hongtao Zhu and Nathan

Hodges for their support and help with the Hille 100 rolling mill.

Thanks to Prof Ernst Kozeschnik from the TU Vienna for his guidance, patience

and help with the thermo-kinetic modelling.

iv

Furthermore, I want to express my gratitude to my collaborators Prof Simon Ringer

and Dr Mehdi Eizadjou and the technical support by Dr Takanori Sato at the

University of Sydney who enabled me to use their atom probe facilities.

I thank my fellow group mates for the stimulating discussions, help with

experimental work and fun we had at our social activities. I especially want to

acknowledge Christian Legerer and Cerys Edwards who cast the steels used in

Chapter 7. A big thanks to my master by coursework student Hafsah Pratiwi for

her hard work and the fun we had working in the lab together.

I want to acknowledge the voestalpine Stahl Linz GmbH (Austria) for supplying the

industrial steel used for the majority of this thesis.

From the bottom of my heart I want to thank my friends who have been my family

away from home. Special thanks to Caitlin who has been my go-to person for all

questions regarding the English language and living in Australia. I want to thank

Arslan for his assistance with the TEM and Christina for her help with the IVAS

software. A special thanks to Koshy and the MSE Badminton Club for sparking my

passion for badminton and to the “ANZAC day”- lunch group for all those delicious

Asian foods we shared. This has helped me to maintain my work-life balance. I

also want to thank my coffee break buddies Irene and Bernd for their moral support

when PhD life was hard.

Last but not least, I would like to thank my family: my parents and my brother for

always supporting me and my partner Fan who is my source of happiness and

calm.

Table of Contents ________________________________________________________________________________________________________________________________

I

Table of Contents

Acknowledgements ........................................................................................... iii

Index of Abbreviations ..................................................................................... IV

Abstract ............................................................................................................. VI

1 Introduction .............................................................................................. 1

2 Literature Review ..................................................................................... 4

2.1 Steel industry, applications and needs ................................................... 4

2.2 HSLA steels ........................................................................................... 4

2.3 Role of Nb, Mo and Cr as microalloying elements ................................. 6

2.4 Processing of HSLA steels ..................................................................... 9

2.5 advanced TMP ..................................................................................... 13

2.5.1 Warm rolling in the ferrite region ................................................ 13

2.5.2 Tempforming ............................................................................. 15

2.5.3 Cold rolling and annealing ......................................................... 17

2.5.4 Influence of C content, heating rate and starting grain size on the

final microstructure ................................................................................... 20

2.6 Mechanism of grain refinement ............................................................ 22

2.7 HSLA steels and aTMP ........................................................................ 24

2.8 Challenges and drawbacks during aTMP ............................................. 25

2.9 Proposed solution to overcome these problems .................................. 29

3 Methods .................................................................................................. 31

3.1 Steel used in Chapter 4-6 ..................................................................... 31

3.2 Steels used in Chapter 7 ...................................................................... 32

3.3 Thermo-mechanical processing ........................................................... 33

3.3.1 aTMP for Chapter 4 ................................................................... 33

3.3.2 aTMP for Chapter 5 ................................................................... 34

3.3.3 aTMP for Chapter 6 ................................................................... 35

3.3.4 aTMP for Chapter 7 ................................................................... 36

3.4 Simulation approach and set-up for Chapter 5 ..................................... 37

3.5 Microstructural characterisation ........................................................... 39

3.5.1 Light optical and stereo microscopy .......................................... 39

3.5.2 SEM/ECCI/EBSD ...................................................................... 40

3.5.3 TEM ........................................................................................... 40

3.5.4 Atom probe ................................................................................ 41


II

3.6 Mechanical characterisation ................................................................. 42

3.6.1 Hardness testing ........................................................................ 42

3.6.2 Tensile testing ........................................................................... 42

3.6.3 Charpy impact testing ................................................................ 42

4 Feasibility study ..................................................................................... 43

4.1 Introduction .......................................................................................... 43

4.2 Results ................................................................................................. 44

4.2.1 Grain refinement ........................................................................ 44

4.2.2 Microstructure of the shear band ............................................... 48

4.2.3 Detailed characterisation of nanoscale precipitates and

dislocation structures ............................................................................... 51

4.3 Discussion ............................................................................................ 55

4.3.1 Grain refinement during aTMP .................................................. 56

4.3.2 Macroscopic shear band formation and CDRX .......................... 57

4.3.3 Direct ageing ............................................................................. 59

4.3.4 Microstructural model and mechanical properties ..................... 60

4.4 Summary and Outlook .......................................................................... 62

5 Optimisation of advanced thermo-mechanical process design ........ 63

5.1 Introduction .......................................................................................... 63

5.2 Results ................................................................................................. 64

5.2.1 Experimental results .................................................................. 64

5.2.2 Simulation results ...................................................................... 75

5.3 Discussion ............................................................................................ 79

5.3.1 Grain size and texture ............................................................... 79

5.3.2 Precipitates ................................................................................ 82

5.4 Summary and Outlook .......................................................................... 84

6 Initial report on structure-property relationships of HSLA steels

subjected to aTMP ............................................................................................ 86

6.1 Introduction .......................................................................................... 86

6.2 Results ................................................................................................. 87

6.2.1 Microstructure ............................................................................ 87

6.2.2 Mechanical Properties ............................................................... 91

6.3 Discussion ............................................................................................ 97

6.4 Summary and Outlook ........................................................................ 100

7 Influence of Cr/Mo as microalloying elements in HSLA steels

subjected to warm deformation .................................................................... 102


III

7.1 Introduction ........................................................................................ 102

7.2 Results ............................................................................................... 103

7.2.1 Hardness ................................................................................. 103

7.2.2 Dislocations and interfaces ...................................................... 105

7.2.3 Solute clusters and second phase precipitates ........................ 109

7.3 Discussion .......................................................................................... 116

7.3.1 Dislocations and interfaces ...................................................... 116

7.3.2 Precipitates and clusters ......................................................... 117

7.3.3 Microstructural Model .............................................................. 118

7.4 Summary ............................................................................................ 120

8 Conclusion and Outlook ...................................................................... 121

8.1 HSLA steels and aTMP ...................................................................... 121

8.2 Mechanical properties ........................................................................ 122

8.3 Alloy design ........................................................................................ 122

8.4 Recommendations for future research ............................................... 122

9 References ............................................................................................ 124

10 Appendix A: MatCalc script ................................................................ 133

Index of Abbreviations ________________________________________________________________________________________________________________________________

IV

Index of Abbreviations

ADF annular dark field

AP(T) atom probe (tomography)

at% atom percent

(a)TMP (advanced) thermo-mechanical processing

bcc body centred cubic

BF bright field

CDRX continuous dynamic recrystallisation

DBTT ductile-to-brittle transition temperature

DDRX discontinuous dynamic recrystallisation

DRX dynamic recrystallisation

EBSD electron backscatter diffraction

ECCI electron channelling contrast imagining

EDS energy-dispersive x-ray spectroscopy

fcc face centred cubic

GB grain boundary

HAGB high angle grain boundary

HRTEM high-resolution transmission electron microscopy

HSLA high strength low alloy

IPF inverse pole figure

LAGB low angle grain boundary

LOM light optical microscopy

m% mass percent

MS martensite start temperature

ND normal direction

RD rolling direction

Rxx recrystallisation

SEM scanning electron microscope

(S)TEM (scanning) transmission electron microscope

Index of Abbreviations ________________________________________________________________________________________________________________________________

V

TnR ‘non-recrystallisation temperature’ (below this temperature complete

static recrystallisation does not longer occur between deformation

passes or during cooling)

UFG ultrafine grained

UTS ultimate tensile strength

wt% weight percent

α ferrite

ε logarithmic strain

𝜀̇ strain rate

γ austenite

Abstract ________________________________________________________________________________________________________________________________

VI

Abstract

Engineering microstructures in high strength low alloy (HSLA) steels via advanced

thermo-mechanical processing (aTMP) is a promising approach to overcome

challenges around low work hardenability and toughness in ultrafine-grained mild

steels.

Thus, the applicability of advanced thermo-mechanical processing for achieving

multi-scale hierarchical microstructures in a HSLA steel was studied. The

microstructural evolution during warm deformation of a martensitic/bainitic starting

microstructure using a Gleeble 3500 thermo-mechanical simulator at 600°C

followed by a direct ageing step was investigated. A strain rate of 10 s-1 during a

single pass plane strain compression led to strain localisation and, therefore, the

formation of a macroscopic shear band. Hence, an optimised advanced multi-hit

thermo-mechanical process for achieving homogenous hierarchical

microstructures without strain localisation was developed. The success of the

process design was verified via high-resolution microscopy, such as electron

channelling contrast imaging, electron backscatter diffraction and transmission

electron microscopy, and thermo-kinetic modelling, using the software MatCalc. A

typical body-centred-cubic rolling texture is achieved in contrast to the previous

process design. Ultrafine crystallites confined by a mixture of high angle gain and

subgrain boundaries are formed, decorated by two types of precipitates. Large

FeMnC-rich cementite particles are found on grain boundaries and smaller

TiNbC-rich precipitates on dislocations and subgrain boundaries. The further aim

of this study was the investigation of the mechanical properties by upscaling the

process developed in the Gleeble using a Hille 100 rolling mill. As a majority of

studies on advanced thermo-mechanical processing of low-alloyed steels was

performed under large strain conditions (<80% reduction in thickness) or on

subsize specimens there is a need to provide mechanical data from standardised

tensile and Charpy impact tests. It was found that rolling to reduction of thickness

of 55% at a temperature around 650°C can lead to an ultimate tensile strength

(UTS) of 650 MPa, a yield ratio of 0.95 and a total elongation of 14% in the

as-rolled condition. The low yield ratio can be explained by the rather large

Abstract ________________________________________________________________________________________________________________________________

VII

precipitates (average precipitate size for the as-rolled condition: 51±38 nm) which

do not effectively contribute to work hardening. Delaminations did occur in the

lower temperature region of the Charpy impact testing in longitudinal and

transversal directions. However, for the 60 min direct aged sample a very high

impact energy of 197±6 J at room temperature was found due to the onset of grain

growth during recrystallisation. Finally, a similar processing was implemented to

invoke grain refinement coupled with strengthening arising from microalloying. For

example, microalloying with Mo and Cr has been reported to provide significant

additional strengthening during such processing but the detailed mechanisms yet

remain unknown. Therefore, three model Fe-1.6Mn-0.04C-0.1Nb+0.5Mo/Cr steels

were warm-rolled at ~600°C with an overall reduction of 50% followed by ageing.

It was found that Mo leads to an increase in hardness of ~20% compared to the

base alloy, whereas Cr provides only a minor hardening increment. It was found

that Mo is more effective than Cr in delaying dislocation recovery. It was also

observed that Mo partitions into nanoscale Nb-C solute clusters and precipitates

of NbC and Fe3C during ageing, retarding the coarsening of these phases.

However, Cr was found to partition into Fe3C only, and does not contribute to the

nature of the dispersion of clusters and NbC.

Introduction ________________________________________________________________________________________________________________________________

1

1 Introduction

Worldwide, steel markets are expanding due to growing population resulting in

expanding urban communities [1–3]. Therefore, novel higher performance steels

are needed for enabling mechanical design with thinner structural elements,

leading to weight reduction and, thus, reduced fuel consumption and CO2

emissions in transport applications. Due to their exceptional cost-performance

ratio, modern HSLA steels are excellent candidates to cover the corresponding

growing demands for advanced steels for construction and transportation [2,3].

HSLA steels are conventionally manufactured by a sophisticated temperature and

deformation schedule known as thermo-mechanical processing (TMP), which

allows to precisely control the microstructural evolution. Here, engineering of a

designed target microstructures can lead to the desired mechanical properties, i.e.

high strength and elongation. Newer, advanced thermo-mechanical processing

routes which include lowering of rolling temperatures to the ferrite region or

annealing of cold-rolled martensite have been shown to have a significant effect

on grain refinement in mild steels. This leads to grain sizes in the submicron

regime, resulting in exceptional yield strength [4–6]. It has been proposed that the

mechanism behind the grain fragmentation is continuous dynamic

recrystallisation [4,7,8]. However, challenges such as low work-hardening

capability and delamination due to rather large cementite particles still need to be

overcome in such steels [4,9]. It is further known that steels with hierarchical

microstructures, which are defined as microstructures exhibiting features across

several length-scales, are able to overcome the well-known strength-elongation

trade-off [10,11]. Thus, the approach of this thesis is to overcome the issues during

aTMP of mild steels with designing aTMPs for modern HSLA steels. This is

because nano-scale precipitates present in such steels are assumed to more

uniformly decorate dislocation networks, by creating the desired hierarchical

microstructures, avoiding delamination and increasing the work hardening

rate [12]. An additional goal is to explore how far the limits for achieving ultrafine

grains that consist of mostly high angle grain boundaries (HAGBs) (i.e. ratio of

HAGBs to low angle grain boundaries (LAGBs)) can be pushed with processing

Introduction ________________________________________________________________________________________________________________________________

2

methods that can potentially be up-scaled. In this thesis Nb and Mo, as well as

their combined effect as microalloying elements, are studied in detail because Nb

is one of the most common and well-studied microalloying elements in HSLA

steels. The main benefits of Nb are precipitation strengthening and its major role

in solute drag and grain boundary pinning [13–16]. Mo is known to increase

hardenability and can delay phase transformations and recrystallisation [17,18].

Alloying with both, Nb and Mo, can lead to a change in the precipitation sequence

[18] which, so far, is not well understood. Further focus is laid on Cr as

microalloying element because it is known to inhibit cementite coarsening [19,20].

The long-term strategic goal of this research program is to achieve an ultimate

tensile strength of ~1100 MPa and a ductile to brittle transition temperature that

might be lower than ~120°C. This thesis investigates only a part of this research

program aimed on the design and implementation of a novel thermo-mechanical

procedure, initially using a simpler steel composition. It is not expected to achieve

a reduction in forces. However, this new processing route, which is limited to rolling

and sheet material, is expected to be more time and energy efficient in terms of

energy used for heat treatments. The design of hierarchical microstructures is a

common aim in modern process and alloy design. However, the detailed aim of

this thesis is to obtain these in HSLA steels, because superior properties

compared to mild steels subjected to advanced thermo-mechanical processing are

expected.The main project objectives discussed in the following chapters are:

1. Study the feasibility of HSLA steels for aTMP to obtain hierarchical

microstructures.

2. Design of an aTMP route that is optimised for achieving homogenous

microstructures in HSLA steels.

3. Up-scale the process to thoroughly study the mechanical properties and

therefore the success of this approach.

4. Study the effect of Cr and Mo as microalloying elements in a Nb

microalloyed steel that has been warm deformed in the ferrite.

To summarise, the main goal of this thesis is to study the microstructure-property

relationship of HSLA steels that have been subjected to deformation in the ferrite

Introduction ________________________________________________________________________________________________________________________________

3

region with correlative high-resolution microscopy and mechanical testing. This

experimental work is accompanied by thermo-kinetic modelling using the software

package MatCalc.

Literature Review ________________________________________________________________________________________________________________________________

4

2 Literature Review

2.1 Steel industry, applications and needs

Since 1950 there has been a steady growth in crude steel production peaking in

2018 at 1,808 million tonnes [1]. Steel products are very versatile as they combine

strength and formability in a way that is tailored to the specific application.

Moreover, they are 100% recyclable. The largest steel market is building and

infrastructure at about 50% followed by mechanical equipment and the automotive

industry [2]. These expanding steel markets can be explained by the world’s

population growth which is expected to reach a maximum of 2.7 billion people by

2050 leading to fast urbanisation [3]. A reduction in usage of natural resources

and accompanying emissions is vital in enabling a sustainable future. Steelmakers

are providing structural solutions for building smart houses that are energy

efficient. Automotive industries are innovating in light-weight design utilising

thinner structural elements and, thus, leading to reduced fuel consumption and

greenhouse gas emissions. These developments clearly highlight the need for

higher performance steels that meet higher environmental standards [3]. HSLA

steels are excellent candidates to fulfil these needs because of their remarkable

strength to weight ratio.

2.2 HSLA steels

A low C steel with small amounts of (micro-) alloying elements is called high

strength low alloy steel. For the first time in the 1960s small amounts

(0.005-0.03 wt%) of Nb were added to simple mild steels which led to an increased

strength up to 415 MPa with relatively good toughness [21]. In contrast to simpler

steels, such as plain or mild steels, the key mechanical properties of HSLA steels,

hardness and yield strength, are increased significantly by low additions of

microalloying elements. Common microalloying elements are Nb, Mo, V and Ti,

and they are known to form precipitates that provide precipitation hardening and,

more importantly, grain size control via Zener pinning and solute drag during

processing [13,16,21]. Figure 2.1 pictures the strengthening effects of precipitates


5

in ferrite (Figure 2.1a) and the influence of the precipitate size and volume fraction

on the maximum grain size (Figure 2.1b). It can clearly be seen from these

diagrams that the maximum strengthening occurs when the precipitates are small,

finely dispersed and have a high volume fraction [16].

Figure 2.1 - a) Precipitate strengthening in ferrite based on Ashby-Orowan model, b)

Effect of precipitate volume fraction and size on maximum grain size [16].

Furthermore, these elements may act as ferrite solid solution strengtheners in the

finished products. Typical amounts added are <0.25 wt%C, 0.02-0.06 wt%Nb/V,

0-0.3 wt%Mo, 0-0.06 wt%Ti. So-called modern HSLA steels further contain

additional elements such as Ni, Cu and Cr, which are known to increase

toughness, corrosion resistance and slow down precipitate coarsening. The

amount of C for modern low-C HSLA steels is typically kept between

0.03-0.06 wt% in order to provide enhanced toughness and weldability. The loss

in strength by the reduced C content is counterbalanced by the additional alloying

elements [16]. Typical yield strengths of mild steels, HSLA and modern HSLA

steels are ~420 MPa, ~690 MPa and ~850 MPa, respectively [13,16,21]. The most

commonly used alloying elements and their purpose are listed in Table 2.1.

a) b)


6

Table 2.1- Alloying elements used in HSLA steels, adapted from Vervynckt et al. [16].

Element Amount in

HSLA/wt%

Influence

C <0.25 Strengthener

Mn 0.5-2 Delays austenite decomposition during accelerated

cooling

Mild solid solution strengthener

Decreases ductile to brittle transition temperature

Si 0.1-0.5 Deoxidiser in molten steel

Solid solution strengthener

Al >0.02 Deoxidiser

Limits grain growth (AlN)

Nb 0.02-0.06 Very strong ferrite strengthener [Nb(C,N)]

Grain size control

Delays γ to α transformation

Ti 0-0.06 Grain size control (TiN formation)

Strong ferrite strengthener

V 0-0.1 Strong ferrite strengthener (VN)

N <0.012 Forms TiN, VN and AlN

Mo 0-0.3 Promotes bainite formation

Ferrite strengthener

Ni 0-0.5 Increases fracture toughness

Cu 0-0.55 Improves corrosion resistance

Cr 0-1.25 Improves atmospheric corrosion resistance (when Cu is

also added)

2.3 Role of Nb, Mo and Cr as microalloying elements

Nb is one of the most common and well-studied microalloying elements in HSLA

steels. The main benefit of Nb is its strong tendency to form Nb(C,N), which does

not only contribute to precipitation strengthening but also pins grain boundaries.

This can result in finer grain sizes and delayed austenite (γ) to ferrite (α) phase


7

transformation. Furthermore, the solute drag effect of Nb, where solute Nb

segregates to dislocations and grain boundaries, is a major contributor to

increasing strength and retarding recrystallisation in steels [13,14,16,22].

Takahashi et al. [22] were able to directly show Nb segregations to dislocations

using atom probe tomography in a Nb ferritic stainless steel as shown in Figure

2.2.

Figure 2.2 - Atom probe results of the Nb added stainless steel. The maps on the right

are rotated around the axis parallel to the measurement direction (the arrow) by about

70°. a) Nb elemental maps. Nb enriched regions are indicated by the arrowheads. b) 1.2%

isoconcentration surface of Nb atoms in the same volume [22].

The addition of Mo as microalloying element is known to increase hardenability

and to delay phase transformations and recrystallisation [18,23] as seen in Figure

2.3a. However, one of the most beneficial effects of Mo alloying was found in

conjunction with Nb as it plays a significant role in the precipitation sequence as it

reduces the activity of C and N (Figure 2.3b) [18].


8

Figure 2.3 - a) Influence of Mo-Nb alloy combinations on static recrystallisation at high

temperature b) influence of Mo addition on dynamic precipitation behaviour of Nb [18].

So far, the role of Mo in the precipitation process is still not fully understood

although it clearly leads to finer and more densely dispersed precipitates which

are more stable upon ageing. There are reports that Mo (i) reduces the diffusivity

of other carbide forming elements [23,24], that (ii) segregates at the

precipitate/matrix interface between precipitate and matrix [25], (iii) is incorporated

into the NbC composition [26–28] and (iv) acts as nucleation site for NbC [18,25].

Most studies investigating the effects of Mo were conducted on TiMo microalloyed

steels [27,29,30]. Park et al. [31] studied a 0.4Nb-0.2Mo microalloyed steel which

was hot rolled and then aged at temperatures between 500-700°C. They did not

observe Mo within NbC precipitates but rather found it in solid solution in the

matrix.

Nb-C and Mo-C cluster formation was observed by Hara et al. [32] in HSLA steels

that were alloyed with B. As B segregates to the austenite grain boundaries it can

inhibit ferrite transformation and hence lead to improved strength. However, as B

strongly tends to form precipitates this is only possible if B is in solid solution.

Therefore, the decrease in C and N diffusion due to the clustering with Nb or Mo

is beneficial to avoid precipitation of B on the grain boundary as can be seen in

Figure 2.4 [18,32].

a) b)


9

Figure 2.4 - Schematic diagram showing the mechanism for suppression of Fe23(C,B)6

due to combined addition of Nb and B or Mo and B [32].

Cr is usually added to steel due to its ability to enhance corrosion resistance [16].

It is also well-known that Cr added to a low C steel can promote the formation of

Cr-carbides which contribute to strengthening [33–36] and slow down cementite

coarsening due to dual partitioning behaviour of Cr and Mn between cementite

and ferrite matrix [19,20]. However, the use of Cr as microalloying element has

not been frequently studied.

2.4 Processing of HSLA steels

HSLA steels are shaped into semi-finished products via thermo-mechanical

processing where the thermal treatments and deformation passes follow specific

schedules aiming to improve mechanical properties via controlled precipitation,

segregation/clustering, solid solution strengthening, and grain refinement [16]. For

achieving best results, it is crucial to perfectly adjust the processing parameters

including temperature, number of deformation passes, interpass time, strain, strain

rate, and cooling rate depending on the steel chemistry.

Grain size refinement may be regarded as the key strengthening mechanism in

HSLA steels as it can be utilised to increase the yield strength and toughness, two

generally inversely related properties [16]. Following the Hall-Petch relationship,

yield strength increases linearly with the inverse square-root of the average grain

size (i.e.) σy = σi + k/√D, where σi is the intrinsic friction to the movement of


10

dislocations in the lattice, k is a material constant and D is the average grain

size [37]. Traditional thermo-mechanical processing (TMP) where deformation

occurs in the lower austenite temperature region is well established and grain

sizes of around 2-5 μm are generally achieved [38–42]. Here, the grains are

mainly refined via the phase transformation from pancaked austenite grains to fine

ferrite grains. Figure 2.5 shows a schematic diagram of a thermo-mechanical

processing route. Firstly, the selection of the solution annealing temperature is

important. Selecting a solution annealing temperature is a fine trade-off between

getting all microalloying elements in solid solution and avoiding grain coarsening

of the austenite. In Ti-alloyed steels, very stable TiN particles form which cannot

be dissolved during typical solution annealing treatments and, thus, restrict the

austenite grain growth [16,38].

Figure 2.5 - Schematic diagram of thermo-mechanical processing and product

microstructures that result from this process [16]. Tnr = non-recystallisation temperature,

Ar3 = temperature at which austenite to ferrite transformation begins during cooling,

Bs = Bainite-Start-temperature, Ms = Martensite-Start-Temperature F = Ferrite,

P = Pearlite, B = Bainite.


11

The second factor that influences the grain size is the temperature region in which

the final deformation step occurs. Figure 2.6 illustrates the microstructural

evolution during deformation above and below TnR (non-recrystallisation

temperature) [16]. In a classic TMP route, the last deformation steps are

performed in the so-called non-Rxx-region. This is the temperature range where

full static recrystallisation (Rxx) between rolling passes or during cooling is no

longer possible. As a result, highly deformed and elongated grains (so-called

pancaked structure) and deformation bands emerge which consequently act as

nucleation sites for the γ to α transformation as can be seen in Figure 2.6b [16,38].

The third parameter that needs to be adjusted is the cooling rate. Accelerated

cooling results in bainitic or martensitic microstructures which can provide higher

yield strengths [16,38].

Figure 2.6 - a) Illustration of deformation above Tnr with complete static recrystallisation

of austenite between rolling passes i and i+1, b) illustration of deformation below Tnr with

only partial recrystallisation of austenite between rolling passes i and i+1: heavily

deformed (“pancaked”) austenite grains result in more nucleation sites for austenite to

ferrite transformation [16].

a)

b)


12

Isasti et al. [39] used this “classic” TMP successfully to achieve grain refinement

of Nb-Mo microalloyed steels deformed in the non-Rxx region. For a

0.05C-1.57Mn-0.05Si-0.028Nb-0.31Mo-0.028Al-0.005N steel at a coiling

temperature of 450°C, a mean unit size (which considers low angle (2-15°) and

high angle (>15°) misorientations) of 2.4 µm was achieved. Kim et al. [40]

investigated the influence of different deformation and coiling temperatures on a

Fe-0.075C-0.2Si-1.7Mn-0.035Al-0.175Ni-0.16Cr-0.275Mo-0.17Ti-0.005N steel.

The processing schedules can be seen in Figure 2.7. Figure 2.7a shows rolling in

the γ Rxx region at temperatures between 1150°C-1050°C. The processing route

for rolling in the γ non-Rxx region at temperatures between 980°C-880°C is

depicted in Figure 2.7b. The average grain size for the deformation in the

recrystallisation region was ~7 μm whereas for the non-Rxx region it was ~3 μm.

Figure 2.7 - Schematic illustrations describing the TMCP schedules: a) γ recrystallisation

schedule (FRT 1050 °C) and b) γ non/recrystallisation schedule (FRT 880 °C), 7 passes

with total thickness reduction of 90% [40].

Further lowering of the processing temperatures in the α+γ two-phase region, as

has been shown by Kim et al. [41], leads to improved impact toughness by

providing a bimodal grain size distribution. During rolling in the two-phase region

austenite can transform into ferrite which is further deformed and can

subsequently undergo recrystallisation which leads to refined grains. The

non-transformed austenite is heavily deformed (pancaked) and transforms to

ferrite upon cooling. However, the finest ferrite grain size (~1.9 μm) and highest


13

strength was be achieved by rolling in the non-Rxx region [41]. Similar results were

reported by Ghosh and Mula [42] who studied a Nb-Ti microalloyed steel at

different deformation temperatures (in the Rxx region, non-Rxx region and the α+γ

region) and for different coiling temperatures. The best balance between the

mechanical properties was be achieved with deformation in the two-phase region

because of the bimodal grain size distribution. Fine grains (~2 µm) are favourable

for strength and larger grains (~35 µm) for ductility.

2.5 advanced TMP

More recent attempts in thermo-mechanical processing aim to achieve ultrafine

grained (UFG) microstructures with grain sizes below 1 μm. This is because

exceptional mechanical properties with yield strengths of 1,000 MPa and beyond

are unlocked in this regime. It is widely known that severe plastic deformation is a

suitable method to achieve these UFG microstructures although the high

logarithmic true strains (up to 5.6) required and the small material volumes limit

the feasibility of such approaches in industrial-scale processing [4,43,44]. More

industrially viable approaches that only use comparably small logarithmic true

strains, typically in the range from 1-3.6, are so-called advanced

thermo-mechanical processing routes [4]. Examples are warm deformation in the

ferrite region of various starting microstructures such as ferrite/pearlite or

martensite, tempforming (deformation of tempered martensite) [4,5,9,45–50] or

intercritical annealing of cold-rolled martensite [6,51–59].

2.5.1 Warm rolling in the ferrite region

Warm deformation describes processing that occurs in the ferrite region which has

been shown to be successful in producing ultrafine grains by several research

groups. The processing route used by Song et al. [48] for an Fe-0.22C-0.21Si-

0.74Mn-0.004P-0.003S-0.029Al-0.001N steel is shown in Figure 2.8. The steel

was austenitised at 920°C for 3 min followed by air cooling to 870°C where a first

deformation step (ε= 0.3, 𝜀̇= 10 s-1) occurred to achieve a completely

recrystallised austenite. After accelerated cooling to the pearlite finish temperature

of 550°C and holding for 2 min, plane strain compression in four passes (each

pass ε= 0.4, 𝜀̇= 10 s-1) was applied. As a final step, an annealing treatment at


14

550°C for 2 h was performed. These authors reported that a critical strain of 0.8 is

necessary to achieve sufficient grain refinement. The grains of the ferritic-pearlitic

starting microstructure elongate, and the cementite lamellae orient perpendicular

to the deformation direction after strain is applied. Some cementite lamellae

fragment into small pieces which are located on the grain boundaries (GB) of the

pearlitic ferrite. These pieces transform into spheroidal particles at large strain

(particle size 90-350 nm). Inside the ferrite grains, also some smaller cementite

particles were found (particle size ~5-90 nm) [48].

Figure 2.8 - Processing schedule for plane strain warm compression tests to achieve

ultrafine grains, Trh: reheating temperature, Ar3: austenite to ferrite transformation

temperature, Pf: pearlite finish temperature [48].

Calcagnotto et al. [49] used a similar processing route, with deformation at 550°C.

However, they applied an intercritical annealing step after deformation to achieve

a ferritic/martensitic finish microstructure. They achieved ultrafine grain sizes of

around 1 µm with a UTS of 633 MPa, a yield ratio of 0.91 and a total elongation of

13.3%. Through warm deformation of a low C steel with a martensitic starting

microstructure, Qun et al. [50] produced UFGs (grain sizes of 1-2 μm). The

specimens were austenitised at 900°C for 20 min and then quenched. Warm

deformation was implemented at temperatures between 600-700°C, strains up to


15

0.7 and strain rates of 0.001-1 s-1, followed by water quenching. The

corresponding mechanical properties can be seen in Figure 2.9.

Figure 2.9 - Stress-strain curve of the low C steel without and with warm deformation at

temperatures of 600°C, 650°C, 700°C under strain rates of 0.001 s-1 and a strain

of 0.7 [50].

2.5.2 Tempforming

Tempforming refers to warm deformation of a tempered martensitic

microstructure. Ohmori et al. [5] were able to achieve UFG for Fe-0.1-to-0.3C-

0.31Si-1.5Mn-0.01P-0.001S steels with varying C-content subjected to

tempforming. The TMP route is depicted in Figure 2.10. Samples were

austenitised at 1100°C, quenched and then tempered at 620°C for 1 h leading to

tempered martensite. After a multi-pass warm calibre-rolling with varying rolling

temperatures (TR) from 400°C-650°C and a reduction in thickness of 93%

(22 rolling passes with reheating after a few passes), the samples were quenched.

A final annealing at temperatures between TR-50 K and TR+100 K for 1 h was

carried out.


16

Figure 2.10 - Thermo-mechanical processing route to obtain UFG low C steel. Pre-heat

treatment followed by a multi-pass warm calibre-rolling with varying rolling temperatures

(TR) from 400°C-650°C and a reduction in thickness of 93% and water quenching. Final

annealing at temperatures between TR-50 K and TR+100 K for 1 h [5].

The microstructure consisted of ferrite grains and spheroidised cementite

particles. The average grain sizes for each rolling and annealing temperature can

be seen in Figure 2.11. It ranges from 0.4 μm for the as-rolled condition at low

deformation temperatures to 16 μm for the highest rolling and annealing

temperatures.

Kimura et al. [45,46] attained a UFG microstructure in a Fe-0.6C-2Si-1Cr and a

Fe-0.4C-2Si-1Cr-1Mo steel which were tempformed at 500°C with an applied

strain of ~1.8. The samples were then air-cooled and the final microstructure

consisted of a UFG fibrous ferrite grains with spheroidal cementite particles

distributed inside the grains. The samples showed a strong <110>‖ rolling

direction (RD) fibre texture which was beneficial for impact properties as through

lamellate fracture the cracks were branched. The phenomenon of delamination

toughening will be discussed in more detail in Chapter 2.8. The average transverse

grain sizes of the elongated grains was 0.37 μm. Lee et al. [47] investigated a

Fe-0.4C-1.96Mn steel. After austenitising at 850°C for 1800 s and quenching the

samples were tempered at 650°C and promptly warm deformed from 10% – 70%

(strain rate 1.1x10-4 s-1). They achieved an equiaxed grain structure with a grain

size of ~2 μm and uniformly distributed cementite particles. Grain refinement


17

through warm deformation of a martensitic starting microstructure has been shown

to be not only successful for steels but also for Ti-6Al-4V alloys [60–62].

Figure 2.11 - Nominal ferrite grain sizes observed on the cross section of samples rolled

at different temperatures (400°C-650°C) and with varying C contents (0.1-0.3 m%) [5].

2.5.3 Cold rolling and annealing

Some authors [63,64] have shown that cold-rolling of a Fe-0.13C-0.0043N-0.01Si-

0.37Mn-0.02P-0.004S steel with a martensitic microstructure and subsequent

annealing can lead to UFG in the nanometre regime. The samples were

austenitised at 1000°C for 1.8 ks followed by water quenching to achieve a

martensitic starting microstructure. Cold-rolling with a reduction in thickness of

50% (equivalent strain of 0.8) occurred in 3 passes in a two-high mill followed by

annealing at temperatures from 200°C to 700°C for 1.8 ks. The microstructure

after annealing consisted of equiaxed UFG with an average grain size of 180 nm

with some parts of block-like morphology and uniformly dispersed nano-carbides.

Annealing above 600°C led to grain and precipitate coarsening [63]. The as-rolled

specimen reached a tensile strength of 1.5 GPa but the elongation was limited to

only 1%. The best ratio of high strength and sufficient elongation were achieved

when annealing occurred at temperatures between 500°C-550°C (see Figure 2.12)

[63].


18

Figure 2.12 - Nominal stress-strain curves a Fe-0.13C-0.0043N-0.01Si-0.37Mn-0.02P-

0.004S steel 50% cold-rolled and annealed at various temperatures for 1.8 ks. Starting

microstructure was martensite [63].

Grain refinement of a low C steel alloyed with Nb was achieved by

Abbasi et al. [54] through cold rolling to a reduction of thickness of 85% of a

martensitic starting microstructure. Grain sizes of 79 nm - 87 nm were produced

after annealing at 550°C for 300 s.

Lan et al. [51] achieved ferrite grains in the nanometre region (200-500 nm) in

high-Si low C and a high-Si microalloyed steel through cold rolling (50% reduction

in thickness) and tempering (450°C-650°C for 60 min) of a martensitic starting

microstructure. Ghassemali et al. [56] reduced the grain size of a plain low C steel

to the UFG range (~188 nm) through cold rolling of an aged martensitic starting

microstructure. They reported that fragmentation of the martensite during cold

rolling plays an important role for the Rxx. Hosseini et al. [58] achieved a ferrite

grain size of 142 nm in a low C steel after 75% reduction in thickness and

annealing at 500°C for 65 min (ultimate strength of 1135 MPa and uniform

elongation of 11.6%) of a martensitic mild steel. Several authors [6,53,57] reported

to have obtained similar grain refinements with a dual-phase starting

microstructure (ferrite/martensite). The applied strain is heterogeneously

distributed in these steels because the deformation concentrates on the soft ferrite

matrix whereas the hard martensite is only marginally deformed.

Azizi-Alizamini et al. [57] achieved a bimodal ferrite grain size structure (<2 μm

and 3-15 μm) by 50% cold rolling and annealing of a dual-phase starting

microstructure of a low C steel. This bimodal structure resulted in similar tensile


19

strength (~550 MPa) and elongation (~12%) compared to a cold-rolled fully

martensite but no Lüders strain was detected.

The aTMP route Mazaheri et al. [53] used for a 0.17C-0.4Si-1.15Mn-0.95Cr-

0.035S-0.025P steel is presented in Figure 2.13. Annealing and quenching were

used to achieve a martensitic starting microstructure. To avoid possible cracking

of the martensite during cold rolling an intercritical annealing step was applied

leading to a ferritic-martensitic microstructure before deformation. Cold-rolled

sheets with a reduction of thickness between 50-80% were annealed at 600°C for

either 20 min or 40 min and then water quenched.

Figure 2.13 - Thermo-mechanical processing route to produce ultrafine/nano ferrite-

carbide microstructures, CR = cold rolling, WQ = water quenched [53].

Figure 2.14 shows the corresponding scanning electron microscopy (SEM)

micrographs for different stages of the processing route. The obtained average

ferrite grain size is 0.35 μm with a few coarser grains of 1 μm and the cementite

particles were ~140 nm (on the grain boundaries) and ~70 nm (inside the grains).

Related mechanical properties can be seen in Figure 2.15. 80% cold-rolling and

annealing at 600°C for 20 min led to a UTS of 880 MPa and a total elongation of

13%.


20

Figure 2.14 - SEM microstructures of: a) as-quenched, b) intercritically annealed, c) 80%

cold-rolled, d) & e) annealed at 600°C for 20 min; F = ferrite, M = martensite, C = carbide,

RD = rolling direction, ND = normal direction [53].

Figure 2.15 - Engineering stress-strain curves of steel specimens obtained by various

thermo-mechanical conditions [53].

2.5.4 Influence of C content, heating rate and starting grain size on the final microstructure

It was found that the C content [5,55,65] as well as the applied heating rate [66]

play a significant role on the final microstructure obtained for steels that have been

subjected to aTMP.


21

Ohmori et al. [5] reported that generally the ferrite grain size during tempforming

decreases with increasing C-content for a Fe-0.1-to-0.3C-0.31Si-1.5Mn-0.01P-

0.001S steel. The same phenomenon was observed by Poorganji et al. [65] for a

tempformed Fe-2 m%Mn-C alloy with varying C-contents between 0.1 and

0.8 m%. They suggest that with increasing C-content the block width of the starting

microstructure decreases whereas an increase in cementite particles was

observed which suppressed dynamic recrystallization (DRX). Foroozmehr et al.

[55] also investigated the influence of the C content on microstructure of a cold-

rolled and annealed low C steel. They found that the martensite becomes finer

with higher C content and the necessary amount of cold reduction for gaining UFG

decreases as well.

Furuhara et al. [66] showed that the heating rate of martensitic steel has a large

influence on the cementite precipitation behaviour (see Figure 2.16). On the one

hand, the nucleation rate increases with increasing temperature and, on the other

hand the size of precipitations is smaller because it takes a longer time to reach

the precipitation nose (less time for Ostwald ripening). Therefore, it is beneficial to

apply a large heating rate to gain a large volume fraction of fine precipitates.

Figure 2.16 - Schematic time-temperature-precipitation diagram describing the effect of

heating rate on cementite precipitation [66].

Li et al. [67] investigated the influence of the initial martensite grain size and pre-

tempering before warm compression at 600°C (strain rate of 1.7·10-3 s-1) on the


22

Rxx behaviour of a SM490 martensite steel. To gain martensite with different grain

sizes, austenitisation was conducted at different temperatures. These authors

found that the Rxx grain size (2.3 μm) only depends marginally on the starting grain

size. Interestingly, there is a similar dynamic recrystallisation (DRX) behaviour for

the as-quenched and pre-tempered samples although cementite particles already

existed after pre-tempering. This suggests that the cementite precipitates

achieved in this material are not effective in pinning grain boundaries.

2.6 Mechanism of grain refinement

Most reports on the mechanism behind the grain refinement during advanced

thermo-mechanical processing suggest that it is in-situ Rxx also called continuous

dynamic recrystallisation (CDRX) [4,7,8,47,52,55].

According to Gourdet et al. [68] the rate of recovery is so high that instead of a

classic recrystallization behaviour continuous dynamic recrystallisation occurs in

high stacking fault materials such as ferrite. Dislocations introduced during

deformation form subgrains, which are confined by low angle grain boundaries,

and then progressively increase their misorientation angle until they convert into

high angle grain boundaries. Low angle grain boundaries usually develop at low

strains and upon reaching strains around 1, a conversion into high angle grain

boundaries can be observed. These microstructures typically consist of a mixture

between crystallites confined either by low angle grain boundaries (LAGBs) or high

angle grain boundaries (HAGBs) [68]. In the following chapters, the concept of

grain size is based on this definition. A schematic drawing of a CDRX

microstructure is shown in Figure 2.17.


23

Figure 2.17– Schematic representation of the CDRX microstructure. Low angle grain

boundaries (thin lines) with a certain misorientation angles θ1-3 are formed due to a steady

accumulation of dislocations (increase in dislocation density ρi). After reaching a critical

value they convert into high angle grain boundaries (thick lines). D is the crystallite size

which is approximately the initial subgrain size [68].

Eghbali et al. [7] reported that with increasing strain the grain size decreases

through formation of subgrains by grain subdivision during CDRX for a low C

microalloyed steel. Furthermore, they observed a significant influence of the strain

rate. The higher the strain rate, the finer the grains as the time for grain growth is

limited [7].

Lan et al. [51] suggest that in martensitic starting microstructures that are cold-

rolled and then aged, two main mechanisms are responsible for grain refinement.

Firstly, the microstructure is subdivided due to its martensitic nature.

A martensite itself is already a fine-grained structure as it consists of numerous

HAGBs (block boundary, packet boundary and prior γ boundary) and lath LAGBs

(see Figure 2.18) [66]. Therefore, they enhance inhomogeneous deformation

which leads to grain subdivision. Furthermore, they can act as nucleation sites for

ferrite grains and precipitates [64,66].


24

Figure 2.18 - Schematic illustration describing potential nucleation sites in martensite [66].

Secondly, the generation of dislocation cell blocks during cold deformation plays

an important role. During the heat treatment, the dislocations rearrange and,

therefore, sharp grain boundaries can be formed. These mechanisms were also

observed by Malekjani et al. [52] in an cold-rolled Nb microalloyed 350 grade steel

and by Foroozmehr et al. [55] in a low C steel. Foroozmehr et al. [55] also pointed

out that the high dislocation density leads to inhomogeneous deformation during

straining and, therefore, to recovery which results in UFGs during annealing.

2.7 HSLA steels and aTMP

So far, most research exploring the potential of aTMP routes was done using mild

steels. Only a few studies focus on microalloyed steels, although several research

groups reported that cold rolling and annealing may be a suitable method to

achieve ultrafine grain sizes in HSLA steels [6,51,52,54,55]. This is why I want to

explore other aTMP routes such as warm deformation in the ferrite on HSLA

steels. Cheng et al. [8] explored the influence of different processing routes for a

Ti-Mo microalloyed steel starting with rolling in the austenite and finish rolling in

the ferrite followed by direct ageing at 600°C for 30 min, achieving final grain sizes

of 1 μm. The resulting UTS was 898 MPa with an elongation of 14.7%.

Gallego et al. [69] performed warm deformation at 740°C of a water-quenched


25

HSLA steel which led to a final grain size of 0.9 μm and a combination of cementite

particles with a size of 114.7±45.5 nm as well as fine (Nb,Ti)(C,N).

2.8 Challenges and drawbacks during aTMP

Low work hardening rates due to relatively large cementite particles and

delamination are common problems in mild steels that have been subjected to

aTMP and still need to be addressed [4,9]. It has been reported that the increase

in yield strength achieved by ultrafine grain sizes has to be paid off by a reduction

of elongation which, subsequently, leading to delamination and a reduced upper

shelf energy as seen in Figure 2.19 [4,9,70].

Figure 2.19 - Dependence of the Charpy impact properties on temperature of steels with

different ferrite grain sizes. The symbol dα refers to average ferrite grain diameter.

Ductile-to-brittle transition temperature of subsize specimen (DBTTsubsize) with a 1 mm

notch depth and a ligament size of 3 mm x 3 mm [9].

It has also been suggested that the rolling texture and elongated grains play an

important role in the delamination phenomenon [4,9]. Song et al. [9] reported that

ultrafine grains in a C-Mn steel leads to increased strength compared to a coarser

grain size, however ductility decreases as seen in Figure 2.20. They suggested

that during tensile testing at room temperature, the refined microstructure lowered


26

the amount of work hardening due to the strong dynamic recovery in body centred

cubic metals. It is assumed that there are two recovery mechanisms occurring

simultaneously. Firstly, slow recovery inside the grains and secondly faster

recovery at grain boundaries. In UFG metals the time needed for dislocations

moving to grain boundaries is shorter than available during tensile testing.

Therefore, recovery at the grain boundaries is already initiated which reduces the

dislocation density inside the grains and simultaneously the work hardening

effect [9]. It is suggested that work hardening increases with the amount of finely

dispersed precipitations inside the grains due to generation of geometrically

necessary dislocations [9,12].

Figure 2.20 - Engineering stress-strain curves of steels with different ferrite grain sizes.

The different grain sizes were produced by conventional route (without large warm

deformation) and UFG route respectively. The UFG route involved a warm deformation

procedure with four steps (each step with ε =0.4, �̇�= 10 s-1) and a subsequent 2 h

annealing treatment at 823 K. The symbol dα refers to the average ferrite grain

diameter [9].

Furthermore, Song et al. [4,9] reported a more pronounced delamination in the

investigated steels with decreasing grain size or Charpy impact testing

temperature. The reason for this delamination could not be clarified as there are

many possibilities like texture phenomena, precipitates on grain boundaries (GBs)

or distorted ferrite-pearlite microstructures. Figure 2.21 depicts the fracture


27

surface of a Charpy testing sample as well as a cross-section scanning electron

microscopy (SEM)/ electron backscatter diffraction (EBSD) investigation. In Figure

2.21b decohesion of GBs is indicated by the smooth undulating surface.

Delamination along GBs with different texture components in the rolling direction

is shown in Figure 2.21c and d.

Figure 2.21 - SEM micrographs and ND (normal direction) orientation map (taken by

electron backscatter diffraction (EBSD) measurement) of ultrafine-grained 0.2% C steel

(average ferrite grain diameter of 1.3 mm) after subsize Charpy impact testing at 103 K.

The images shown in (a) and (b) are taken from a plane normal to the rolling direction

(RD), while (c) and (d) are normal to the transverse direction (TD) of the sample.

Orientation components in (d), <111>‖ND in blue, <001>‖ND in red and <101>‖ND in

green. (a) overall fracture surface: (b) transition between delaminated and shear fracture

regions. The observation area of (b) is shown in (a). (c) longitudinal cross-section. The

black arrows point out chains of large voids in the specimen; (d) crack propagation along

interfaces. The circles 1 and 2 show two elongated grains with high-angle grain

boundaries in between. The observation area of (d) is shown in (c) [4,9].


28

Controversially, there have been studies showing that delamination can be

beneficial to increase low-temperature toughness and decrease the ductile-to-

brittle transition temperature. Usually, there are two types of laminate geometries

(see Figure 2.22): (i) the crack divider orientation where a crack splits up into

multiple cracks at the weak interface resulting in a change in stress state, (ii) in

the crack arrester orientation, a crack re-initiation is necessary due to the blunting

of the crack tip as a result of the change in stress state [70].

Figure 2.22 - Schematic illustration of laminate geometry showing delamination

toughening [70].

Frequently, it has been reported that in ultrafine-grained steels deformed by

multipass calibre rolling delamination occurs due to the formation elongated

microstructures [9,71–73]. Kimura et al. [74] and Min et al. [75] reported that

delamination can lead to increased toughening in high C steels which have been

calibre rolled. These steels exhibited a distinct <110> // RD fibre texture and

therefore cracks are branching in longitudinal direction leading to an inverse

temperature dependence of impact toughness.


29

2.9 Proposed solution to overcome these problems

In this thesis, I aim to achieve a first step towards tackling the issues discussed in

section 2.8 via engineering hierarchical microstructures in a modern HSLA steel.

My approach is to achieve UFG microstructures by warm deformation of

martensitic/bainitic starting materials in the ferrite. I further propose controlled

precipitation and segregation of alloying elements to dislocations, subgrain and

grain boundaries through a direct ageing step following deformation. I hypothesize

that nm-sized dispersed precipitates are more beneficial for work hardening

because of their interaction with geometrically necessary dislocations [12] and that

ductility will be increased due to the absence of large carbides on grain

boundaries.

Figure 2.23 shows a schematic of the target microstructures that I want to achieve

in HSLA steels that have been subjected to aTMP and the microscopy techniques

that I will be using to reveal these microstructural features. In Chapter 4 I will be

exploring the feasibility of using an industrially produced HSLA steels to obtain

hierarchical microstructures with aTMP. Chapter 5 focuses on the optimisation of

the aTMP design for this industrial steel used in this thesis in order to obtain

homogeneous microstructures. The complete optimised processing route is

modeled to reveal the complexity of particle evolution, including precipitation

kinetics during different processing stages as well as the grain size evolution. The

success of the process design will be evaluated in Chapter Error! Reference

source not found. by up-scaling this process using a Hille 100 rolling mill and

mechanical testing such as tensile and Charpy impact testing. Finally, in Chapter

7 I will make a first step towards an alloy design for HSLA steels that are optimised

for the processing route developed. Three steels will be cast and rolled on a lab

scale to systematically study the influence of common alloying elements Nb, Mo

and Cr during warm deformation in the ferrite. I will use high-resolution microscopy

techniques to study the microstructures obtained. Electron channeling contrast

imaging (ECCI) as well as EBSD are used to reveal the grain size and grain

boundary character. TEM is utilised to study the change in morphology evolution

of nano-scale precipitates whereas atom probe (AP) is a powerful tool to reveal

the chemical composition on an atomic level and therefore enables to uncover

segregations and clusters.


30

Figure 2.23 – Schematic of the target microstructure in HSLA steels which have been

warm deformed in the ferrite and microscopy techniques used to reveal microstructural

features.

Experimental Methods ________________________________________________________________________________________________________________________________

31

3 Methods

In this chapter the steels used for this study are presented. For Chapter 4-6 a

commercially available steel was utilised whereas for Chapter 7 three lab-scale

steels were produced using arc-melting. The thermo-mechanical processing

routes for each chapter are discussed separately. Additionally, all microstructural

and mechanical characterisation methods and the parameters applied are

reported in detail.

3.1 Steel used in Chapter 4-6

The steel used for the majority of this study (Chapter 4-6) is a commercially

available modern HSLA steel produced via continuous casting and its chemical

composition is given in Table 3.1.

Table 3.1 - Chemical composition of the modern HSLA steel in this study, bal. is Fe.

C Si Mn Nb Cr Ni Al Cu Ti Mo N

wt

% 0.047 0.06 1.92 0.045 0.165 0.01 0.052 0.015 0.016 0.113 0.0054

A 180 tonne test melt was produced via an industrial-scale processing route at the

voestalpine Linz, Austria. The material was produced in a blast furnace, followed

by secondary metallurgy in a ladle including the RH (Ruhrstahl Heraeus) process

to remove dissolved gases. The steel was then continuous cast. Figure 3.1 shows

a schematic of the dimensions of the slab and the plates that were cut from it. The

slab (~25 tonne) had the dimensions of 215 mm x 1330 mm x ~11 m (H x W x L).

Plates with the dimensions of 260 mm x 215 mm x 20 mm (W x H x D) were cut

from the edge of this slab in order to avoid the middle segregation and to ensure

homogeneous composition across all samples. The microstructure of the plate

was dendritic ferrite. Smaller pieces were cut from this plate and wrapped in

stainless steel foil to prevent oxidation during the following heat treatment.

Solution annealing was done in a muffle furnace at 950°C for 10 min (30 min for

the samples in Chapter 6) followed by water quenching.


32

Figure 3.1 - Sketch of the industrial slab where the initial plates were cut from the edge.

3.2 Steels used in Chapter 7

Three steels were fabricated from raw materials (pure forms of Fe, C, Mn, Nb, Mo

and Cr) using an Edmund Buehler AM200 vacuum arc melting furnace. After

melting, the samples had dimensions of ~100 mm length and an elliptical cross-

section with diameters of ~14 mm x 9 mm. They were homogenised in a tube

furnace under argon atmosphere at 1100°C for 10 hrs followed by furnace cooling.

The resulting chemistry was determined by inductive coupled plasma mass

spectroscopy and can be seen in Table 3.2. These three steels will be denoted as

Nb, Nb-Mo and Nb-Cr in Chapter 7.

Table 3.2 - Steel compositions in wt.% of three model HSLA steels made by vacuum arc

melting as determined by inductively coupled plasma spectroscopy.

Fe Mn C Nb Mo Cr other

elements (S,

P, Cu, Ni, Si,

V, Ti, Al and B)

Nb bal. 1.58 0.033 0.1 < 0.01 < 0.01 ≤ 0.01

Nb-Mo bal. 1.61 0.041 0.1 0.53 0.01 ≤ 0.01

Nb-Cr bal. 1.6 0.038 0.1 < 0.01 0.51 ≤ 0.01


33

3.3 Thermo-mechanical processing

3.3.1 aTMP for Chapter 4

The deformation and direct ageing experiments were conducted in a Gleeble 3500

thermo-mechanical simulator in plane-strain compression mode under rough

vacuum (~1x10-1 Pa). Standard plane-strain compression samples with the

dimensions of 10 mm x 15 mm x 20 mm were utilised. Ta-C foils on the samples

as well as Ni paste on the anvils were used to guarantee a proper contact between

sample and anvils and to reduce friction. The temperature was monitored with a

type K thermocouple which was spot welded to the sample surface. A schematic

of the aTMP is shown in Figure 3.2. The martensitic/bainitic starting materials were

heated to a deformation temperature of 600°C with a heating rate of 10 K/s

followed by a 120 s soaking time to obtain trough-heating of the samples. Different

logarithmic true strains from 0.2 to 1 were applied in a one hit deformation pass

with a constant strain rate between 0.01-10 s-1. The majority of experiments was

conducted with the strain rate of 10 s-1, which was chosen in order to be close to

typical industrial-scale rolling speeds. The direct ageing was carried out in the

Gleeble directly after deformation at 600°C and the holding time was varied

between 0-60 min (0 min means no ageing was conducted). Following aTMP, all

samples were air cooled. The results from this processing route are discussed in

Chapter 4.


34

Figure 3.2 - Schematic of the proposed aTMP processing schedule for achieveing

multiscale-hierarchical microstructures in modern HSLA steels.


Deformation experiments were performed using a Gleeble 3500 thermo-

mechanical processing simulator in plane-strain compression mode. Standard

plane-strain compression samples with the dimensions of 10 mm x 15 mm x 20

mm were utilised. A K type thermocouple was utilised to monitor the temperature

during the experiments. Cu paste on the anvils was used to ensure good contact

between samples and anvils. In this study, deformation pass numbers varied

between 2 and up to 6 passes with varying strains between 0.1 and 0.2 per pass.

The strain rate as well as the deformation temperature were kept constant at 1 s-

1 and 873.15 K/600°C, respectively. Two different soaking times prior to

deformation, 2 min and 10 min, as well as two different interpass times, 30 s and

60 s, were investigated to optimise our previous process design. Following

deformation, direct ageing times were applied between 1-60 min. Figure 3.3 shows


35

the final optimised aTMP route, which resulted in the most homogenous

microstructures. The results of this processing route are discussed in Chapter 5.

Figure 3.3 - Optimised aTMP route with 600 s soaking time before deformation at 600°C

with a strain rate of 1 s-1, 3 hit deformation with 0.2 strain each and 30 s interpass time,

followed by direct ageing times of 1-60 min.


The same industrially produced Fe-0.05C-1.9Mn-0.05Nb-0.1Mo-0.016Ti steel as

described in Chapter 3.1 was used. Samples with dimensions of 200 mm length,

50 mm width and 20 mm thickness and a tapered end were wrapped in a stainless

steel foil to prevent oxidation during solution annealing. This was carried out in a

muffle furnace at 950°C for 40 min followed by water quenching. Tapering was

necessary to guarantee a smooth feeding of the samples into the rolls. The

samples were reheated to 650°C in a muffle furnace for 30 min to enable a

through-warming and then rolled in a Hille 100 rolling mill in a two-high setting with

a roll diameter of 220 mm. The rolling was conducted in 3 passes to achieve an

overall reduction on thickness of 55%, as can be seen in Table 3.3. The roll speed

was 40 rpm and the roll gab was set to 14 mm, 10 mm and 8 mm for each pass,


36

respectively. The samples were reheated for 2 min in-between passes,

immediately direct aged with times varying from 0 min – 60 min after the final pass

and subsequently air cooled. The surface temperature before and after each pass

was monitored using an infrared thermometer.

Table 3.3 - Thickness reductions and surface temperatures per rolling pass.

thickness surface temperature after each

pass

initial 20 mm ~650°C

pass 1 15.5 mm ~620°C

pass 2 11 mm ~590°C

pass 3 9 mm ~570°C


Samples of the steels described in Chapter 3.2 were cut into slices with a thickness

of 2 mm, length of ~14 mm, a width of ~9 mm and subsequently subjected to a

simplified thermo-mechanical processing route as shown in Figure 3.4a. Solution

annealing at 950°C for 10 min was done using a muffle furnace followed by water

quenching. The samples were wrapped in stainless steel foil to reduce oxidation

and decarburisation during the heat treatment. For warm rolling, all samples were

reheated to 600°C in a Carbolite furnace between two copper blocks for 5 min and

then rolled in a customised mini-rolling mill (see Figure 3.4b) which has a roll

diameter of 64.5 mm. An overall reduction in thickness of 50% was achieved in

10 passes with a reduction of 0.1 mm during each pass resulting in a final

thickness of 1 mm. All samples were placed in the Carbolite furnace after each

pass for ~1.5 min to maintain the deformation temperature of 600°C. Finally,

selected samples were aged in a muffle furnace at 600°C with varying times from

10 min – 60 min and then air cooled. The results of this processing are discussed

in Chapter 7.


37

Figure 3.4 - a) Schematic thermo-mechanical processing schedule for deformation and

ageing of model HSLA steels in the ferrite, b) customised mini-rolling mill with a roll

diameter of 64.5 mm.

3.4 Simulation approach and set-up for Chapter 5

The software package Matcalc 6.02 (rel 0.014) was used to simulate the thermo-

kinetics of the precipitation and grain size evolution of the processing route shown

in Figure 3.3. The precipitate evolution in this software is based on an efficient

mean-field description and the thermodynamic extremum principle as described

by J. Svoboda, F.D Fischer, P. Fratzl and E. Kozeschnik (also known as SFFK

model) [76,77]. This model can provide insights in the evolution of precipitates with

focus on multi-particle multi-phase multi-component nucleation, growth and

coarsening, chemical composition and the influence on the stability of the

precipitates, taking into account the misfit between precipitate and matrix. The

grain size was modelled according to a state parameter-based model developed

by Buken et al. [78,79] to accurately model simultaneous precipitation and

recrystallisation as well as the competing process of recovery. An initial equilibrium

simulation with the original chemistry of the alloy system was performed to predict

possible phases in this steel. AlN, NbC, TiN, Cementite, M7C3 were suggested. A

T0-temperature calculation was used to obtain a solidus temperature at 1478°C

and a martensite start temperature. The T0-temperature is defined as the

temperature at which two phases with an identical chemical composition have the

same molar Gibbs free energy and therefore can quantify diffusionless phase

transformations. A preliminary simulation with the Scheil-Gulliver simulation


38

module of MatCalc (taking back diffusion of C and N into account) [80,81] has

shown that only approximately 40% of Ti and N will be bound in primary TiN

precipitates and 60% of the nominal Ti and N are remaining in solid solution after

solidification, when performing the Scheil-Gulliver simulation under the

assumption that final solidification occurs at 3% residual liquid fraction. This means

that most of the Ti and N are available for precipitation in the matrix during the

later heat treatment. Furthermore, the martensite start temperature was calculated

as 492°C, in order to switch the precipitation domains, austenite and ferrite,

according to temperature changes. Thus, proper diffusion coefficients of alloying

elements in austenite and ferrite are taken into account.

Following preliminary calculations, it was decided to simplify the system only

allowing alloying elements and precipitates that were found in the actual

experiments. The system was simulated following the complete processing route

including casting, solution annealing at 950°C for 10 min and quenching, followed

by the aTMP route according to Figure 3.3 with deformation occurring in the ferrite.

The initial computing parameters used in the simplified simulation are listed in

Table 3.4. A large initial subgrain diameter was chosen as the simulation starts

from casting. In the standard setup, MatCalc automatically creates a complex

carbo-nitride if Ti, Nb, N and C are selected to form a precipitate with face-centred-

cubic (fcc) crystal structure. However, this complex carbo-nitride shows a

miscibility gap. Therefore, the carbo-nitride phase is split into two phases, one with

Nb and C as the major constituents and a second one where Ti and N are major

constituents. Furthermore, grain boundaries were selected as nucleation sites for

cementite, whereas nucleation sites for TiN and NbC were set as dislocations and

subgrain boundaries, as found in experimental observations. Results of these

simulations are shown and discussed in Chapter 5.


39

Table 3.4 - Computing parameters used for simulation, all other parameters are used as

suggested by default values within the present MatCalc version 6.02.0014.

Calculated solvus temperature 1478°C

Calculated martensite start 492°C

Elements used in simplified

calculation

Fe, Mn, Nb, Ti, N

Phases used in simplified

calculation

fcc_A1, bcc_a2, cementite

Major constituents for

fcc_a1#01

Nb, C

Major constituents for

fcc_a1#02

Ti, N

Ferrite-subgrain evolution-

model

one-parameter ABC [82]

ABC parameters (A/B/C) 50/2/1e-4

Similitude parameter A’ 40

Initial ferrite grain diameter 100e-6 m

Initial subgrain diameter 100e-6 m

Nucleation sites NbC

(fcc_a1#01)

dislocations & subgrain-

boundaries

Nucleation sites TiN

(fcc_a1#02)

dislocations & subgrain-

boundaries

Nucleation sites cementite grain-boundaries

3.5 Microstructural characterisation

3.5.1 Light optical and stereo microscopy

Specimens for stereomicroscopic imaging and LOM were ground and polished to

1 µm and subsequently etched with Nital (5% HNO3 in ethanol). The imaging was

done using a LEICA M205A (stereomicroscope) and a Nikon ME600 LOM

equipped with a DSRI2 camera.


40

3.5.2 SEM/ECCI/EBSD

After standard grinding and polishing to 1 µm finish samples for electron

channelling contrast imaging (ECCI) and EBSD were electropolished using a

STRUERS Lectropol-5 apparatus and a STRUERS electrolyte A2 (perchloric acid

in ethanol) at 25 V for 20 s with a flow rate of 10 at room temperature. ECCI was

carried out using a ZEISS Auriga field emission SEM at 20 kV, 9 mm working

distance, and 30 μm aperture size with no rotation or tilt of the stage.

EBSD was done in a JEOL 7001 field emission SEM equipped with an EDAX

Hikari Super EBSD system at 25 kV, 15 mm working distance, 6x6 binning and a

50 nm step size (a step size of 0.5 μm was used for imaging of the starting

material). The TSL OIM Analysis 7 software was used for EBSD data evaluation.

An EBSD data-clean up with a neighbour confidence index correlation of 0.1 was

applied prior to grain/subgrain size evaluations with a grain tolerance angle of 2°,

and a minimum size 3 data points as threshold values. For grain size

measurements, an average of 3-5 EBSD scans was used, high angle grain

boundaries (HAGB) were defined with an angle >15°, whereas subgrain

boundaries, i.e. low angle grain boundaries (LAGB) were set between 2-15°. The

grain size was always defined as crystallites that are either confined by HAGBs or

LAGBs.

The texture in Chapter 5 was analysed using the TSL OIM Analysis 8 software

with merged data-sets of 3-4 individual scans within the deformation zone. This

was done to obtain more statistically significant information about the overall

texture. However, it must be noted that the homogenous area is small due to the

nature of the plane strain compression test applied in this research.

The fracture surfaces in Chapter 6 were characterised by photography and using

a ZEISS Auriga field emission SEM with secondary electron contrast at 15 kV and

an aperture of 30 µm.

3.5.3 TEM

Slices with a thickness of around 150 µm were cut from the samples and then

punched to discs with 3 mm diameter. These discs were mechanically thinned to

around 60 μm, followed by electropolishing using a STRUERS Tenupol-5 twin-jet

polishing machine and a STRUERS electrolyte A2 (perchloric acid in ethanol).


41

Parameters used were 26 V with a pump flow rate of 15 and a light stop value of

17 at ~7°C.

Transmission electron microscopy (TEM) imaging in Chapter 4 was done using a

Philips CM200 field emission TEM equipped with a Bruker QUANTAX energy-

dispersive x-ray spectroscopy (EDS) system at 200 kV.

The TEM used in Chapter 5-7 was a JEOL JEM-F200 cold field emission gun

(CFEG) scanning TEM (STEM) operated at 200 kV acceleration voltage. Images

were acquired in scanning mode at a camera length of 120 mm with a

corresponding inner angle of 62 mrad collection for the annular dark field (ADF)

detector and 10 mrad for the bright field (BF) detector. This configuration provides

strong diffraction contrast in the BF-STEM and Z contrast in ADF-STEM images.

The microscope is equipped with a JEOL 100mm2 silicon-drift X-ray detector with

a collection solid angle of approximately 1 sr which was used to acquire the EDS

maps.

3.5.4 Atom probe

Atom probe (AP) samples were prepared electrolytically using perchloric-acid at

15-20 V DC. The AP measurements were conducted in voltage mode with a LEAP

3000 Si with a detector efficiency of 0.57 from Cameca at a temperature of 40 K,

a pulse fraction of 20% and a pulse repetition of 200 kHz. The software package

IVAS 3.6.6 from Cameca was used for data reconstruction. Calibration of the

reconstruction parameters such as field factor and image compression factor was

done using crystallographic features present in the atom probe dataset [83].

A cluster analysis was performed after removing the poles from the data with an

iso-surface for Fe at 75%, based on the of the detector efficiency multiplied by the

density of body-centred-cubic Fe = 84.6 atoms/nm3. The total number of atoms

after removing the poles in the data sets shown in Figure 7.9 was 11.3 million and

27.7 million for the Nb-Mo 30 min and Nb-Cr 10 min respectively. This analysis is

based on the maximum separation method [83] and was performed with the first

nearest neighbour distances considering C, C2, C3, and Nb.


42

3.6 Mechanical characterisation

3.6.1 Hardness testing

Vickers hardness testing according to ASTM E92-17 using a STRUERS Duramin

300 (HV5) was done in order to be able to achieve a first estimate of the

mechanical properties. Each value present is an average from at least 5 indents

(10 indents for Chapter 7).

For Chapter 5, a 1 kg load was chosen to allow for sufficient distance for 3 indents

in the homogenous region of the deformed samples.

3.6.2 Tensile testing

Flat dog-bone tensile testing samples with a thickness of 3 mm, a width of 12.5 mm

and a gauge length of 50 mm were machined according to ASTM E8/E8M-11

standard parallel to RD. The tests were conducted with a constant crosshead

speed of 1 mm/min at room temperature using an Instron 5982 tensile testing

machine and an MTS laser extensometer Lx500. Unfortunately, due to issues with

the laser extensometer, the lower yield strength for the 60 min samples is not

available. The 10 min samples started necking outside the gauge length, hence,

these values are greyed out in Table 6.2, but their total elongation after fracture

fitted to the overall trend. Generally, a minimum of 2 samples was used to obtain

lower yield strength, UTS and total elongation for each condition.

3.6.3 Charpy impact testing

Subsize V-notch Charpy samples were machined according to ASTM E23-07a

with 55 mm length and 10 x 7.5 mm2 cross-section along and transverse the rolling

directions. Impact test were performed in a temperature range between -90°C and

120°C using Mohr & Federhaff A.G. pendulum impact testing machine. The

temperature was monitored with a K-type thermocouple that was spot welded onto

the samples. An average of 2-3 samples was used to obtain the impact energy for

each condition.

Feasability Study ________________________________________________________________________________________________________________________________

43

4 Feasibility study

The work in this chapter has been published:

C. Ledermueller, H. Li, S. Primig, Engineering Hierarchical Microstructures via

Advanced Thermo-Mechanical Processing of a Modern HSLA Steel, Metall. Mater.

Trans. A 49, 6337–6350 (2018).

C. Ledermueller did the majority of the planning and execution of the experimental

work as well as the data analysis. The manuscript was drafted by her.

4.1 Introduction

Advanced thermo-mechanical processing has been shown to be successful in

achieving ultrafine grain sizes in steels. However, the majority of studies have

been published on plain mild-steels where problems such as delamination and low

work-hardening rate have been reported.

The advanced thermo-mechanical processing route studied in this chapter is

described in Chapter 3.3.1 and consists of warm deformation of a martensitic

starting microstructure. As martensite consists of packets and blocks which are

HAGBs they can act as nucleation sites for either precipitations or the ferrite

grains [64,66]. As rolling forces are rather high for cold-rolling of martensite a

deformation in the “warm”- ferrite region is proposed. A direct ageing step which

simulates coiling is expected to lead to precipitation of Nb(C,N) at the subgrain

boundaries which stabilises the UFG microstructure and simultaneously leads to

work hardening. This chapter focuses particularly on the feasibility of

accomplishing the target microstructures with such an aTMP using a HSLA steel.

We use a Gleeble thermo-mechanical processing simulator and high-resolution

materials characterisation (see Chapter 3.5.1-3.5.3) to reveal microstructures

achieved. The steel used in this chapter is described in Chapter 3.1. The first

section of the results focuses on grain refinement, the formation of a macroscopic

shear band in most samples is shown in the second section, and the third section


44

reveals the decorations of grain and subgrain boundaries as well as dislocations

with nanoscale precipitates.

Following chapters in this thesis will focus on optimising the aTMP route and up-

scaling for thorough mechanical testing.

4.2 Results

4.2.1 Grain refinement

Figure 4.1a shows the lath shaped microstructure of the starting material after

solution annealing and quenching, which may be described as a mixture of

martensite and bainite. The lath width is approximately 10±5 μm. Figure 4.1b is an

EBSD inverse pole figure map with high angle grain boundaries (HAGB) shown

as black lines and low angle grain boundaries (LAGB) as red lines. Individual

sheaves of bainite and/or martensite can be seen. Figure 4.1c is a TEM bright field

image showing microtwins in the martensite, marked by arrow. The hardness of

solution annealed and water quenched material with 202±4 HV5 is not

considerably higher compared to the received as-cast material with 195±3 HV5.


45

Figure 4.1 - Microstructural characterisation of the as-quenched starting microstructure

for deformation experiments: (a) LOM shows the quite coarse lath-shaped microstructure,

(b) EBSD inverse pole figure map shows individual sheaves, high angle grain boundaries

are shown with black lines and low angle boundaries with red lines, IPF map is from top

view sample has as no real coordinate system yet, (c) The TEM image shows microtwins

in the martensite which are marked with a red arrow.

Figure 4.2 shows the as-deformed microstructure of a sample that was deformed

to a true strain of 0.8 at 600°C without ageing. Overall, grain refinement has been

achieved in the deformed section of the sample, however, strain unfortunately did

not occur in an entirely homogenous manner. Figure 4.2a-c are LOM images at

different magnifications of the cross-section. The blue square in Figure 4.2a

indicates where all further imaging was done in the following sections of this paper,

and the red arrows mark the deformation direction. A macro-shear band originates

from the edges of the Gleeble anvils and crosses the samples diagonally from top

left corner to bottom right corner in this case. The shear band exhibits a very

fine-grained microstructure, although it was not possible to reveal the grains in the

LOM, whereas the microstructure next to the shear band remains somewhat

coarser. Figure 4.2c shows local hardness measurements. The hardness in the


46

shear band is roughly constant around 299±4 HV5 but decreases with increasing

distance to the shear band to 226 HV5 at a distance of 1.2 mm from the centre of

the shear band. According to LOM images the onset of shear band formation

occurred between an applied strain of 0.2 and 0.4. Figure 4.2e shows the evolution

of the hardness in the shear band over true strain for 30 min ageing time and

Figure 4.2f over ageing time at a true strain of 0.8, respectively. Please refer to

Chapter 3.6.1 for a detailed description of the hardness testing. It was found that

hardness increases with an increase in strain and reaches 300±3 HV when a true

strain of 1 is applied, whereas it only drops slightly with an increase in ageing time

to 280±5 HV at a constant true strain of 0.8 and 60 min ageing. Further, it can be

seen that the hardness reaches a plateau after 30 min and does not decrease

further. Figure 4.2d is an exemplary result for lower strain rates applied, in this

case for a strain rate of 0.01 s-1, where it can be seen that these samples still

exhibit a shear band. The further focus of this paper was laid on the higher strain

rate of 10 s-1 as this is closer to industrial rolling conditions.


47

Figure 4.2 - As deformed sample at 600°C, true strain 0.8, strain rate of 10 s-1 and no

ageing time: (a) macroscopic image of the sample which shows the macroscopic shear

band. The red arrows indicate the compression direction and the blue square indicates

the position at which all the following imaging was done, (b) light optical image with a

close up of the microstructure in the centre of the sample, (c) hardness mapping revealing

a constant hardness value in the shear band and a decrease in hardness with increasing

distance from the shear band, (d) exemplary result with lower strain rate of 0.01 s-1 at

same temperature of 600°C, true strain of 0.8 and no ageing time, (e) hardness evolution

as a function of true strain with a constant ageing time of 30 min, (f) hardness evolution

as a function of ageing time for a constant true strain of 0.8.


48

The corresponding Gleeble flow curves show work hardening up to a peak stress

of around 670 MPa followed by continuous flow softening as shown in Figure 4.3.

The individual curves match very closely although it can be seen that the true

strains achieved are slightly lower compared to the targeted true strains. This is

due to machine stiffness and elastic deformation influences. Figure 4.3b shows

the time versus temperature during deformation with adiabatic heating at ~180 s

of around 50°C during deformation at the strain rate of 10 s-1.

Figure 4.3 - Flow curves during plane strain compression tests of the modern HSLA steel

with a martensitic starting microstructure at 600°C with a strain rate of 10 s-1 (a) flow

curves with different true strains, (b) adiabatic heating during deformation.

4.2.2 Microstructure of the shear band

As the deformation was localised in the shear band and resulted in most severe

grain refinement in this area, further focus of this work was laid on analysing the

microstructure in the shear band. Figure 4.4 shows EBSD inverse pole figure maps

with high angle grain boundaries shown in black and low angle grain boundaries

in white. The evolution of the microstructure with increasing strain and a constant

ageing time of 30 min is presented here. For the corresponding IPF colour code,

see Figure 4.1. In Figure 4.4, it can clearly be seen that the initially coarse grains

become elongated in shear direction and subdivide into smaller units with

increasing strain. A true strain of 1 presumably leads to a conversion of most low

angle grain boundaries (LAGB)s or subgrain boundaries into high angle grain

boundaries (HAGB)s. Figure 4.4f shows the evolution of misorientation angle over

true strain. It can be seen that the misorientation of the subgrains increases with

plastic strain (-)

pla

sti

c s

tre

ss (

MP

a)


49

increasing strain. The average misorientation increased from 7.2° for a true strain

of 0.2 to 17.2° for a true strain of 1. No pronounced rolling texture formation in the

shear band was observed. However, the microstructure in Figure 4.4b (true strain

of 0.4), after the onset of the formation of the macroscopic shear band, shows the

typical α-/γ-fibre rolling texture of body-centred cubic metals.

Figure 4.5 shows the evolution of microstructure with a constant true strain of 0.8

with increasing ageing time following deformation. ECCI is a technique that images

crystallites (grains as well as subgrains) but the imaging of the detailed grain

boundary character is not straightforward. This micrograph reveals that grain and

subgrain sizes are in the submicron regime. Furthermore, the grain size of

selected ECCI images was determined manually with linear intercept method to

compare it to EBSD data for EBSD grain size evaluation calibration. The grain size

of the sample with a true strain of 0.8, ageing time of 10 min and a strain rate of

10 s-1 was 0.32±0.17 μm and 0.38±0.03 μm, determined via linear intercept

method and EBSD, respectively. Henceforth, in this paper, the term ‘grain size’

will be used to describe all crystallites (subgrains and grains) confined by either

LAGBs or HAGBs [68]. EBSD further revealed that it is possible to achieve a grain

size of 370±20 nm in the shear band of the as-deformed sample during

deformation to a true strain of 0.8. Conversely, in as-deformed areas outside of

the shear band, an inhomogeneous distribution of grain size is present. It was

found that the UFG structure is stable upon ageing and that crystallites do not

significantly change with increasing ageing time as shown in Figure 4.8a. It

reaches a value of 460±20 nm after 60 min ageing. The average grain size value

of the sample annealed for 30 min is somewhat coarser than the 60 min sample.

Although an average of 5 EBSD scans was used to determine the grain size the

location of the scans might influence these results as the deformation was

inhomogeneous.


50

Figure 4.4 - Evolution of the microstructure of a modern HSLA steel with increasing strain

and constant ageing time of 30 min: (a) true strain = 0.2, (b) true strain = 0.4, (c) true

strain = 0.6, (d) true strain = 0.8, (e) true strain = 1, black lines show HAGBs whereas

white lines denote LAGBs. Large initial grains subdivide into smaller units and with a strain

of 1 a large amount of LAGBs has converted into HAGBs. Please refer to Figure 4.1 for

the inverse pole figure colour code. (f) evolution of misorientation angle over true strain.


51

Figure 4.5 - Microstructural evolution of samples deformed at 600°C with a true strain of

0.8, a strain rate of 10 s-1 and varying ageing times from 0 min to 60 min studied by ECCI,

EBSD and TEM. (a)-(c) 0 min ageing time, (d)-(f) 30 min ageing time, (g)-(i) 60 min time.

Please refer to Figure 4.1 for the inverse pole figure colour code. The red arrows in the

TEM micrographs indicate precipitates.

4.2.3 Detailed characterisation of nanoscale precipitates and dislocation structures

All studies of dislocation structures and precipitates have been done in the shear

band in the centre of the samples. Figure 4.6 shows TEM bright field images of

dislocation structures found in the modern HSLA steel deformed at 600°C to a true

strain of 0.8 with a strain rate of 10 s-1. Figure 4.6a shows recrystallised ultrafine

grains in a sample aged for 30 min. The arrow highlights the overlapping of two

recrystallised grains. Figure 4.6b shows the sample in the 60 min condition where

the fragmentation of grains as well as dislocation pile ups can be observed.


52

Figure 4.6 - TEM images of samples deformed at 600°C with a true strain of 0.8, (a) 30 min

ageing, overlap of two recrystallised grains (marked by red arrow), (b) 60 min ageing,

grain fragmentation, dislocation pile up (marked by red arrows).

Two populations of precipitates are observed as can be seen in Figure 4.7. Larger

precipitates with diameters ranging from 106±49 nm for 0 min ageing time to

190±111 nm for 60 min ageing time are preferably located on grain boundaries

(marked by arrows in Figure 4.7c-d). Smaller precipitates have diameters ranging

from 20±13 nm for 0 min ageing time (see next paragraph for explanation of larger

average value of 0 min sample) and 12±11 nm for 60 min ageing time which also

frequently nucleate on subgrain boundaries (marked by arrows in Figure 4.7a-b)

and individual dislocations. Figure 4.7a-b show bright and dark field images of an

as-deformed sample at a deformation temperature of 600°C with a true strain of

0.8, respectively. Figure 4.7c-d show bright and dark field images for a sample

deformed at 600°C with a true strain of 0.8 and a direct ageing time of 30 min.


53

Figure 4.7 - Two populations of precipitates present after aTMP, studied by TEM bright

field and corresponding dark field imaging: (a)-(b): small precipitates are observed at

dislocations here shown for the 0 min ageing time condition, (c)-(d): large precipitates are

seen at grain boundaries here shown for the 30 min ageing time condition.

Figure 4.8b shows the evolution of precipitate size with increasing ageing time

from 0 min to 60 min which suggests a complex particle size distribution with

particles forming at different stages of the processing route. A total of 227 particles

was analysed for 30 min condition in the range from 3 nm to 500 nm with an

average particle size (small and large particles combined) of 68 nm. A total of 204

particles was analysed for 60 min condition in the range from 3 nm to 430 nm with

an average size of 93 nm. The evaluation for 0 min condition was difficult because

of the small length scale and strain contrast around particles. Therefore, it was


54

only possible to analyse 49 particles in the range from 4 nm to 230 nm with the

same number of images used as for all conditions. Also, most of these precipitates

were of the larger type, sitting on the grain boundaries. Therefore, the average

precipitate size of 74 nm likely overestimates the real value. However, it can be

concluded that the average precipitate size increases from ~70 nm to ~100 nm.

Figure 4.8 - The influence of increase in ageing time from 0 min to 60 min of samples

deformed at 600°C to a true strain of 0.8 and a strain rate of 10 s-1, (a) on the evolution of

the grain size, (b) on the evolution of precipitate size.

Figure 4.9 shows typical EDS spectra and mappings collected in the TEM, carried

out in order to reveal the chemical composition of the precipitates. Here, three

types of precipitates were identified according to their chemistry, namely FeMnC-

rich, CrNiSi-rich and TiNbMo-rich. The larger precipitate type which preferably sit

on the grain boundaries, were identified as FeMnC-rich, whereas the smaller type,

which nucleate on subgrain boundaries and dislocations, is TiNbMo-rich. Also a

single CrNiSi-rich precipitate was found.


55

Figure 4.9 - EDS in the TEM of precipitates found in a HSLA steel after deformation at

600°C to a true strain of 0.8 with a strain rate of 1 s-1 and varying direct ageing times. (a)

30 min ageing time. FeMnC rich and NbMoTi-rich precipitates, (b) corresponding EDS

spectrum of position 1 in (a), (c) corresponding EDS spectrum of position 2 in (a), (d)-(i)

corresponding EDS mappings, (j) 60 min ageing time, CrNiSi-rich precipitate, (k)

corresponding EDS spectrum.

4.3 Discussion

In the following discussion, the individual sections will focus on features of the

hierarchical microstructure achieved. The first section will focus on grain

refinement, and the second section on decoration of grain and subgrain

boundaries as well as dislocations with precipitates. Finally, a schematic model of

the hierarchical microstructure achieved will be presented and discussed.


56

4.3.1 Grain refinement during aTMP

Through warm deformation of a martensitic/bainitic starting microstructure of a

modern HSLA steel, UFG microstructures with grain sizes in the range of

0.37 μm – 0.58 μm were achieved in the shear band (Figure 4.5). The grain sizes

in literature are more homogenous if rolling was used instead of plane strain

compression. Malekjani et al. [52] reported a similar subgrain size of 300 nm for a

Nb microalloyed steel grade 350 (0.145C-0.023Nb) after 50% cold rolling and

subsequent annealing at 500°C between 300-7200 s of a martensitic

microstructure. The applied rolling reduction of 50% which equals a true strain of

0.7 matches the strains studied in our research. The (sub)grain size is similar to

our results with grain sizes in the range of 0.37 μm – 0.58 μm (Figure 4.8a). These

authors found Fe3C on grain and subgrain boundaries as well as on triple

junctions. The average precipitate size was reported to be 67 nm after 300 s and

100 nm after 7200 s annealing. This is slightly smaller compared to the cementite

size measured here which was 106±49 nm for the 0 min ageing condition (Figure

4.8b). Note that the C content of 0.145% was double the amount compared to the

current material (see Table 3.1). We identified the same preferable nucleation sites

for cementite (Figure 4.7c-d). Malekjani et al. [52] further reported large

recrystallised grains of 1.6±0.3 μm after 7200 s annealing, in contrast to our study.

Okitsu et al. [6] achieved grain sizes of 0.49 μm for cold rolling to 91% of a

ferritic/martensitic duplex microstructure (0.1C-0.018Nb) followed by annealing at

625°C for 120 s. They reported that their grains were uniaxial and did not consist

of substructures which they explained by the large reduction in thickness applied.

The grain size is again comparable to our results although these authors used a

larger reduction in thickness of 90% which equals a true strain of 1.6. Furthermore,

they used a martensitic/ferritic duplex starting microstructure where most of the

deformation was located in the softer ferrite. Abbasi et al. [54] achieved UFG sizes

through cold rolling to 85% reduction in thickness of a martensitic microstructure

(0.14C-0.045Nb) and subsequent annealing. Annealing for 300 s at 550°C and

600°C led to grain sizes of 79 nm and 167 nm, respectively. This is somewhat

smaller than in our research which might be due to the larger rolling reduction of

85%, equalling a true strain of 1.9. They furthermore discussed the positive effect

of NbC to prevent grain growth during annealing. However, there is no detailed


57

characterisation of the nano-scale precipitates in their paper.

Foroozmehr et al. [55] studied the influence of C content and rolling reduction on

the formation of UFG of a Nb microalloyed steel. They concluded that the lower

the C content is the higher the required reduction. Up to 80% cold rolling and

annealing at 500°C for 30 min was not sufficient to achieve complete grain

refinement in a 0.038C-0.02Nb steel. They concluded that steels with a C content

lower than 0.08 wt% are not suitable for achieving UFG with this process.

Other methods to achieve UFG in microalloyed steels have been reported

although the resulting grain sizes are slightly larger compared to cold rolling and

annealing. Cheng et al. [8] achieved a grain size of 1 μm in a Ti-Mo microalloyed

steel (0.11C-0.3Mo-0.11Ti) after rolling at 850°C with a reduction of thickness of

55.6%, finish rolling at 650°C with a reduction of 65% followed by ageing at 600°C

for 30 min. They report that strain induced nano-scale (Ti, Mo)C precipitate in the

substructures of the pancaked austenite which is not only beneficial for grain

refinement but for precipitation hardening as well. Gallego et al. [69] reported a

grain size of 0.9 μm in a microalloyed steel (0.16C-0.048Nb-0.016Ti) after warm

deformation at 740°C with a total deformation of 60% of a martensitic

microstructure. They furthermore reported a cementite size of 115±46 nm. It

appears that the grain sizes achieved with these methods are on average larger

compared to our processing approach, although the rolling reductions used are

comparable to the current study.

To conclude, warm rolling in the ferrite of a martensitic starting microstructure of

HSLA steels is a suitable method of aTMP for grain refinement in the submicron

regime. The clear industrial advantage of this method compared to cold rolling and

annealing is the reduced rolling force needed [50]. The advantage compared to

conventional TMP is the finer grain sizes achieved.

4.3.2 Macroscopic shear band formation and CDRX

Although target microstructures have been achieved, inhomogeneous

deformation due to formation of a macroscopic shear band as seen in Figure 4.2

occurs if a true strain of 0.4 or higher was applied. The shear band starts to form

at a strain of around 0.3. A possible reason is adiabatic heating during deformation

with high strain rate of 10 s-1 or the occurrence of gradient temperature prior to


58

deformation. Furthermore, it is belived that the geometry of the samples and anvils

in the Gleeble during plane strain compression lead to strain localisation as

described with finite element modelling by Aksenov et al. [84]. These authors

showed that large strain and strain rate inhomogeneities occur over the volume of

a HC420LA steel for deformation temperatures between 980°C – 1050°C. They

reported that the local effective strain in the sample compared to the nominal strain

applied might be up to 60% higher [84]. The formation of adiabatic shear bands in

steels is well known for high strain rates achieved through ballistic impacts. Xu et

al. [85] reported that shear localisation in a low C steel is dependent on the

strength at a certain strain rate. A higher strength leads to an increased likelihood

of shear banding. They also stated that the critical strain needed for shear band

formation in a quenched steel is 0.3, similar to our results. A similar behaviour of

shear banding was observed by Chao et al. [61] in a Ti-6Al-4V alloy when a

martensitic starting microstructure was uniaxially compressed at 700°C with a true

strain of 0.8 and a strain rate of 1 s-1. These authors suggested that the reason for

the shear band formation could be adiabatic heating which may lead to a phase

transformation and flow softening. They showed that lowering the strain rate to

0.001- 0.01 s-1 led to more homogenous deformation. We too observed adiabatic

heating of 50°C, as can be seen in Figure 4.3b. The approach to overcome shear

banding in this study was using lower strain rates. However, the lowest strain rate

of 0.01s-1 still resulted in shear banding in the material investigated in the current

study as can be seen in Figure 4.2d. Gourdet and Montheillet [68] developed a

model to describe flow curves of high stacking fault energy metals. They

suggested that stress-strain curves show a distinct maximum followed by only

limited softening. Experimental results although revealed a considerably amount

of flow softening. By incorporating adiabatic heating into their model they could

match the experimental observations. These authors concluded that adiabatic

heating and/or other topological effects are responsible for the flow softening.

Therefore, it can be concluded that the flow curve in Figure 4.3a follows typical

flow behaviour commonly observed during CDRX where strain localisation leads

to adiabatic heating (Figure 4.3b). Further, the EBSD scans in Figure 4.4 show the

subdivision of the original coarse grains by LAGBs and the gradual formation of

equiaxed refined grains which are surrounded by HAGBs. Due to their high

stacking fault energy, ferritic steels undergo CDRX, where classical nucleation


59

does not play are role in the formation of new grains. As dynamic recovery is the

predominant mechanism, recrystallised grains form by transformation from

subgrains into new grains within the original grains which have undergone

deformation. Dislocations accumulate in LAGBs leading to an increase in

misorientation angle and finally to the formation of HAGBS when a critical value is

reached [68]. The increase in misorientation angle for our samples is shown in

Figure 4.4f. Gourdet and Montheillet [68] suggest that the formation of HAGBS

happens at moderate true strains of 1. In our research first grains defined by

HAGBS started to form at a true strain of 0.6 (Figure 4.4c), but it can be assumed

that the local equivalent true strains are higher. When a strain of 1 was applied a

majority of subgrains have transformed (Figure 4.4e). This grain fragmentation

was also observed in the TEM, as seen in Figure 4.6, which shows that

dislocations subdivide individual grains. This confirms that the mechanism behind

the grain refinement during warm deformation and direct ageing of martensite is

CDRX, which was also found by other authors [4,7,8,55,61].

4.3.3 Direct ageing

As shown in Figure 4.7, Figure 4.8b and Figure 4.9 two populations of precipitates

were found. Large FeMnC-rich precipitates which were preferably found on grain

boundaries, and smaller precipitates which were NbMoTi-rich and nucleated on

dislocations and subgrain boundaries.

Charleux et al. [86] studied the precipitation behaviour of a Nb and Ti HSLA steel

(0.07C-0.08Nb-0.047Ti) which was processed via traditional TMP and ageing at

650°C for 60 min. They reported spherical Fe3C precipitates with sizes of

0.1-0.5 μm in diameter that formed due to spheroidisation of pearlite colonies

during annealing. As they were relatively large in size, these authors suggested

that they do not contribute significantly to strengthening. They furthermore

observed Nb/TiC of different sizes depending on their location. The precipitate size

on grain or subgrain boundaries ranged from 7-12 nm. Precipitates on dislocations

within grains with high dislocation density were needle-shaped with a length of

3-6 nm and a diameter of 0.7-0.9 nm. Spherical Nb/TiC with a size of 3-5 nm were

observed in grains with a low dislocation density. Lan et al. [51] reported cementite

particles with a size of several nm and NbTiV-rich precipitates with a size of a few


60

nm formed during annealing at 550°C of a cold-rolled martensite (0.16C-0.05Nb-

0.065Ti). These reports match well with the results in our study.

We assume that the larger FeMnC-rich precipitates are cementite which forms

during auto-tempering of the martensite during heating and soaking at deformation

temperature. A similar phenomenon was observed by Li et al. [87] during warm

deformation of a low C martensite with 0.16 wt%C. The steel in the present study

has only 0.05 wt%C which explained the smaller amount of cementite in our

material compared to Li et al.’s study. Other authors reported precipitation of large

cementite particles at higher temperatures or spheroidised pearlite [48,86].

As NbCN, TiN and TiC exhibit a NaCl crystal structure, they have a high lattice

mismatch with the ferrite and precipitate on crystalline defects such as dislocations

or subgrain boundaries. This leads to a strengthening effect following the

Orowan-Ashby mechanism [13]. Therefore, it can be assumed that strain-induced

precipitation occurs during deformation. Misra et al. [88] reported that only smaller

precipitates (<25 nm) are contributing significantly to precipitation hardening.

However, these precipitates do not only contribute to strengthening but also inhibit

grain growth during ageing as can be seen in Figure 4.5 and Figure 4.6a. The

grain size of our steel is stable upon ageing due to Zener pinning of the precipitates

located on grain and subgrain boundaries, as can be seen in Figure 4.7.

Lan et al. [51] discussed these positive effects of microalloying elements upon

ageing for a cold-rolled martensite. They also reported slight coarsening of

precipitates after 60 min tempering. Furthermore, Ostwald ripening during ageing

was observed for all precipitates, as can be seen in Figure 4.8b attributed to the

high diffusion coefficient of Nb and C in ferrite. Due to the chemical composition

of this steel, which also contains other carbide formers such as Cr, it is likely that

the coarsening of the Nb(C, N) is slowed down. It is believed that the isolated

CrNiSi-rich particle has formed because the solution annealing was not sufficient

enough for local homogenous redistribution of alloying elements.

4.3.4 Microstructural model and mechanical properties

In this study we aimed to produce a hierarchical microstructure in a HSLA steel in

order to enhance its mechanical properties by decorating ultrafine grains with

precipitates. A schematic of the hierarchical microstructure achieved is shown in


61

Figure 4.10. Initial coarse grains (in the micrometre regime, bold solid lines in

Figure 4.10) are subdivided by subgrain boundaries which form subgrains in the

nm-regime (dashed lines in Figure 4.10). With increasing strain, these subgrain

boundaries convert into HAGBs (solid lines in Figure 4.10). It was observed that

these grains defined by HAGBs preferably form in a necklace-like structure.

Precipitates (green and purple circles in Figure 4.10) decorate the grain

boundaries and stabilize the microstructure during ageing. It is assumed that this

hierarchical microstructure can provide superior mechanical properties combining

high strength with simultaneously good elongation.

Figure 4.10 - Microstructural model of a hierarchical microstructure achieved in a HSLA

steel via warm deformation of a martensitic/bainitic starting microstructure.

The hardness of the hierarchical microstructure in the shear band which can be

used to estimate the mechanical properties showed an increase of around

100 HV5 compared to the starting microstructure for an applied strain of 1. By

using ISO 18265:2003 [89], table A.1 for unalloyed, low-alloyed steels and cast

iron, a conversion of the hardness values into tensile strengths is possible. The

initial hardness of 200 HV would be a tensile strength of 640 MPa whereas the

highest hardness of 300 HV5 converts into a tensile strength of 965 MPa. The

increase in hardness with increase in true strain is due to a combination of grain

refinement and precipitation strengthening. A slight hardness drop was observed

with an increase in direct ageing time after deformation. Ostwald ripening of the

precipitates was observed in the TEM although grain size did not significantly

increase as can be seen in Figure 4.8. Thus, it is suggested that both precipitate

coarsening and dislocation annihilation are mainly responsible for this hardness

drop. However, the hardness reaches a plateau after 30 min and does not


62

decrease further. This confirms that the microstructures achieved are thermally

stable.

4.4 Summary and Outlook

In this chapter, a stable hierarchical target microstructure was achieved via

advanced thermo-mechanical processing of a modern HSLA steel. Deformation at

600°C with a strain rate of 10 s-1 of martensitic/bainitic starting microstructure

leads to strain localisation and therefore to the formation of a macroscopic shear

band if a true strain of 0.4 and higher is applied. Continuous dynamic

recrystallisation leads to ultrafine grain sizes of around 0.5 μm in the shear band

region which are stable upon ageing at 600°C up to 60 min. Two populations of

precipitates were found, a larger type of FeMnC-rich precipitates which preferably

nucleates at high angle grain boundaries, whereas smaller NbMoTi-rich

precipitates also nucleate at dislocations and subgrain boundaries. These

precipitates undergo limited Ostwald ripening during ageing at 600°C. Such

microstructures are desirable for design of modern HSLA steels with superior

mechanical properties in terms of strength, ductility and work hardening.

Thus, the following chapters will include achieving more homogenous deformation

by modifying the deformation parameters such as number of deformation passes

and strain applied per pass as well as strain-rate and soaking time before

deformation. After successful design approach for homogeneity the process will

be up-scaled in order to be able to enable thorough mechanical testing.

Optimisation of advanced thermo-mechanical process design ________________________________________________________________________________________________________________________________

63

5 Optimisation of advanced thermo-mechanical

process design


C. Ledermueller, E. Kozeschnik, R.F. Webster, S. Primig, Advanced Thermo-

mechanical Process for Homogenous Hierarchical Microstructures in HSLA

Steels, Metall. Mater. Trans. A Phys. Metall. Mater. Sci. 50 (2019) 5800–5815

C. Ledermueller did the majority of the planning and execution of the

experimental work as well as the data analysis. The manuscript was drafted by

her.

5.1 Introduction

In the previous chapter it was shown that hierarchical microstructures in a modern

Nb-Ti-Mo HSLA steel can be achieved via single-hit warm deformation at 600°C

with a strain rate of 10 s-1 and a logarithmic true strain of 0.8 of a

martensitic/bainitic starting microstructure in the ferrite region. We found ultrafine

grains confined by a mix of high angle and subgrain boundaries with an average

crystallite size of ~0.5 μm. This microstructure is decorated by nano-scale

precipitates, larger Fe-Mn rich particles (~150 nm) which preferably sit on high

angle grain boundaries and smaller Ti-Nb rich particles (~15 nm) which nucleate

preferably on dislocation structures within the grain. However, due to the high

strain rate and single pass deformation, these target microstructures occurred only

highly localised, within a macroscopic shear-band.

Thus, the main aim of the current chapter is to optimise the aTMP of the same

steel to achieve more homogenous hierarchical microstructures. This is done via

a systematic variation of the processing parameters, such as number of

deformation passes, reduction per pass, interpass, and soaking times in a Gleeble

as described in Chapter 3.3.2. The success of this approach is verified via

correlative microscopy (see Chapter 3.5.1-3.5.3) and corresponding

thermo-kinetic modelling, using the software MatCalc. The simulation setup is

described in Chapter 3.4. The complete processing route is modelled to reveal the


64

complexity of particle evolution, including precipitation kinetics during different

processing stages as well as the grain size evolution. The complex interplay of

deformation and heat-treatment hinders a thorough through-process experimental

study of the microstructural evolution at any given point during processing.

Therefore, thermo-kinetic modelling is a suitable tool for future optimisation of both

alloy and process design.

The chemical composition of the steel can be found in Chapter 3.1 and the

martensitic/bainitic microstructure after solution annealing and water quenching

can be seen in Figure 4.1.

5.2 Results

5.2.1 Experimental results

5.2.1.1 Process optimisation

Our previous study suggested that a macroscopic shear band started to form

between a true strain of 0.2 and 0.3 during warm deformation at 600°C. Therefore,

in this set of experiments, the maximum strain applied in one deformation pass

was limited to 0.2. In Figure 5.1a-b, two samples with inhomogeneous deformation

are shown: The sample in Figure 5.1a was deformed at 600°C with a strain rate of

1 s-1, six pass deformation with a true strain of 0.1 per pass and 30 s interpass

time, whereas in Figure 5.1b, deformation occurred at 600°C with a strain rate of

1 s-1 and four passes with a true strain of 0.2 each and 60 s interpass time. In both

samples, still, a tendency towards shear band formation exists, as indicated by

individual grains that are elongated diagonally from one edge of the Gleeble tool

to the other, highlighted by orange arrows in Figure 5.1a-b. Figure 5.1c-f depicts

the optimised process that achieved the most homogenous deformation.

Processing parameters were a deformation temperature of 600°C, strain rate of

1 s-1, three deformation passes with a true strain of 0.2 each, 30 s interpass time,

and 10 min soaking prior to deformation. With these processing parameters, a

grain orientation in the high deformation zone that is perpendicular to the loading

direction (indicated by orange arrow) and homogenous is achieved, which can be

seen in Figure 5.1c-d. Figure 5.1e is an ECCI image showing the crystallite

structures obtained with this processing route. Original grains that are confined by

high angle grain boundaries (HAGBs) are subdivided into smaller subgrains


65

confined by low angle grain boundaries (LAGBs). However, it is difficult to

distinguish between these two types of interfaces with ECCI. It can be seen that

the crystallite size is in the submicron regime. In the following, as mentioned

above, 'grain size' will refer to (sub)grain size, which will describe all crystallites

that are either confined by HAGBs or LAGBs. Figure 5.1f is the corresponding

EBSD scan to Figure 5.1c, revealing a typical body-centred cubic (bcc) rolling

texture with <111> and <001> as preferred grain orientations. The inverse pole

figure colouring used denotes orientations parallel to the normal direction (ND,

vertical, direction of loading). A more detailed study of the texture will be presented

in section 5.2.1.3. It can be seen that the <111>//ND orientated grains subdivide

more easily as compared to the <001>//ND grains due to the higher amount of

subgrains formed, i.e. LAGBs indicated as white lines in Figure 5.1f. The ratio of

HAGBs : LAGBs is approximately 1 : 2 in the as-deformed condition and remains

constant upon ageing, according to EBSD. The <001>//ND grains exhibit some

areas of blurry contrast, i.e. not very clearly defined interfaces. Furthermore, the

EBSD scan shows fully recrystallised grains that are confined by black lines, which

indicate HAGBs only. These recrystallised grains have no preferred orientation

and preferably nucleate forming a necklace-like structure. Interestingly, it seems

that the overall strain of 0.6 is sufficient to achieve ultrafine grain sizes. Therefore,

these aTMP parameters (see Figure 3.3) were chosen to study the influence of

direct ageing on the evolution precipitates and hardness.

Figure 5.2a shows the evolution of grain size with increasing ageing time. It can

be seen that the grain size is approximately constant, starting at ~0.5 μm, and

does not significantly change remaining around ~0.55 μm even after being

subjected to direct ageing at 600°C for 60 min.


66

Figure 5.1 - Microstructure of samples at 600°C with a strain rate of 1 s-1. Normal direction

(ND, direction of loading, indicated by 'F') is vertical. (a) 6 x 0.1 strain and 30 s interpass

time, (b) 4 x 0.2 strain and 60 s interpass time, (c) most homogenous deformation,

processing parameters according to Figure 3.3 with 3 x 0.2 strain, 30 s interpass time and

10 min soaking time prior to deformation, (d) light optical micrograph of (c), red square

indicates were microstructural characterisation was done, (e) ECCI of (c) showing

ultrafine grains, and (f) EBSD of (c) showing pronounced bcc rolling texture and grain

refinement. Inverse pole figure colouring indicates directions parallel to ND, white lines

indicate LAGBs, and black lines indicate HAGBs. Orange arrows indicate the grain flow.

The inset shows IPF colouring in normal direction (vertical).

ND

RD


67

5.2.1.2 Hardness

Figure 5.2b depicts the influence of direct ageing on the evolution of hardness.

Please see Chapter 3.6.1 for a detailed description of the hardness testing. The

hardness provided in Figure 5.2b was measured in the sample centre. The grey

band is a 95% confidence band for the fitted trend line and was calculated using

the software Origin. Overall, the hardness appears fairly constant with an average

value of 251±1 HV1 for the as-deformed condition and a hardness of 250±7 HV1

after 60 min ageing. However, the blue trend line suggests a slight decrease of

around 10 HV1 in hardness over time of around -5 HV1.

Figure 5.2 - Evolution of (a) grain size measured by EBSD as crystallites that are

confined either by LAGBs or HAGBs and (b) hardness over ageing time and trend

line.

5.2.1.3 Texture

For texture studies, 3-4 scans were merged to enable a texture analysis over a

larger area. However, the EBSD data is still limited for a significant textural study.

Therefore, inverse pole figures presented in Figure 5.3a-b were mainly used to

confirm the success of the aTMP route. We believe that the typical <111>//ND

and <001>//ND texture found here, which is typical for a bcc material that

undergoes continuous dynamic recrystallisation in the ferrite, is an indicator for

success. However, it has to be mentioned that in a few samples a deviation from

this typical behaviour was found, with one selected example shown in Figure 5.3c.

It is not expected that this texture would change significantly during ageing.


68

Figure 5.3 - Texture with strong <111>//ND and <100>//ND orientation, (a) 1 min ageing

time (b) 60 min ageing time, (c) 10 min ageing time with deviation from typical bcc rolling

texture.

5.2.1.4 Precipitates

Similar to our previous work [90], two types of precipitates are found. Figure 5.4a-b

(0 min ageing) and Figure 5.5a-b (10 min ageing) show bright and corresponding

dark field scanning TEM (STEM) images of the larger (>50 nm), round type. This

threshold value is arbitrarily chosen at this point, but a more detailed analysis of

the precipitate size will be presented in Figure 5.9. These particles preferably

precipitate on grain boundaries (Figure 5.5a-b) but are also frequently found in

areas like the ones depicted in Figure 5.4a-b, with a large number of precipitates

in close proximity. Figure 5.4c-d and Figure 5.5c-d show STEM images of the

smaller (<50 nm) precipitates, which nucleate inside grains, on dislocations and

subgrain boundaries. It can be seen that following deformation most of the smaller

precipitates are initially cuboidal, as shown in Figure 5.4c-d. However, with

increasing ageing time they transform into rather star-shaped morphologies

(Figure 5.5c-d).


69

Figure 5.4 - Bright and dark field STEM images for the 0 min ageing sample: (a) and (b)

large precipitates are found in large quantity in close proximity, (c) and (d) smaller,

cuboidal precipitates, marked by arrows respectively.


70

Figure 5.5 - Bright and dark field STEM images for sample aged for 10 min at 600°C: (a)

and (b) large precipitates preferably nucleate on grain boundaries, (c) and (d) smaller

precipitates become star-shaped.

Figure 5.6 presents EDS mappings of the precipitates. Figure 5.6a shows an area

with a high density of precipitates after 0 min ageing time. It can be clearly seen

that there are two types of precipitates according to their chemical composition:

Larger, round shape precipitates that are rich in Fe, Mn, C and Cr, and smaller,

cuboidal precipitates that are rich in Ti and Nb. Figure 5.6b shows a smaller

precipitate after 10 min ageing that already changed its morphology into star-

shape. It can be seen that the core is Ti, Nb and N rich and depleted in Fe, with

no clear enrichment of C. It also seems that Ti is concentrated in the core of the

precipitate whereas Nb seems to be spread more widely. Figure 5.6c shows the

EDS mapping of a smaller star-shaped precipitate after 60 min ageing. It was

found to be rich in Ti and Nb too, however, no N was detected. Interestingly, C as


71

well as Si form a shell around the precipitate. This mapping shows again that Ti is

concentrated in the precipitate centre whereas Nb is spread more widely.


72

Figure 5.6 - EDS mappings of precipitates in the TEM: (a) 0 min ageing time, showing

larger precipitates that are Fe, Mn and C-rich and smaller precipitates that are Ti and

Nb-rich, (b) 10 min ageing time, small star-shaped precipitate which is rich in Ti, Nb and

N in the core, (c) 60 min ageing time, small star-shaped precipitate which is rich in Ti and

Nb but exhibits a shell of C and Si. N was not detected here, therefore, no mapping is

shown.

To further characterise the precipitates, selected area diffraction was performed

on one of the larger round precipitates in the sample after 60 min ageing (Figure

5.7). The matrix is indexed using a bcc crystal structure with a lattice parameter a

= 0.286 nm. Major spots that belong to the [001] zone axis can be seen in Figure

5.7a. The diffraction pattern of the precipitate, as seen in Figure 5.7c, was indexed

using an orthorhombic crystal structure with cementite lattice parameters a = 0.508

nm, b =0 .673 nm, c = 0.451 nm. It was found that these major spots are the [013]

zone axis and additional spots were indexed as (113) and (222) which means

there is another crystal near [1-32] zone axis. The angle between these two spots

is 45.5°, which matches the angle between these planes. Therefore, it can be

concluded that the precipitate is cementite. From the high-resolution TEM

(HRTEM) image in Figure 5.7d it does not seem that the cementite is coherent

with the matrix. Coherency would mean that the lattice planes of the matrix

continue in the precipitate.


73

Figure 5.7 - Selected area diffraction of larger round precipitates of the sample aged for

60 min. (a) red circle marks the area for the diffraction pattern in (c), (b) diffraction pattern

of the steel matrix, (d) HRTEM of the interface of precipitate and matrix.

Figure 5.8 shows the selected area diffraction of one of the star-shaped

precipitates. The HRTEM, which was taken on one of the arms of the star,

suggests that the arm of TiNbC is (semi-)coherent with the matrix. The diffraction

pattern is difficult to index as only a few spots correspond to the NaCl structure

with a lattice parameter of a = 0.425 nm for bulk NbC. It may be suggested that

some of the spots originate from the arm, and others from the TiN-core of the

precipitate.


74

Figure 5.8 - Selected area diffraction of smaller, star-shaped precipitates of the sample

aged for 60 min. (a) red circle marks the area for the diffraction pattern in (c), (b) diffraction

pattern of the steel matrix, (d) HRTEM of the interface of precipitate and matrix.

In Figure 5.9, the distribution of precipitate size, which was measured from STEM

images, is plotted. The rather complex distribution suggests that particles originate

from different stages of the processing route. For 0 min ageing time 494 particles

were evaluated with an average precipitate diameter of 68 nm, a minimum of 9 nm

and a maximum of 331 nm. For the 10 min sample, 272 precipitates were

evaluated. The average diameter slightly increased to 86 nm with a minimum of

13 nm and a maximum of 631 nm. 445 particles were measured for the 60 min


75

ageing sample resulting in an average value of 113 nm, a minimum of 9 nm and a

maximum of 630 nm.

Figure 5.9 - Precipitate sizes in as-deformed, after 10 min and 60 min direct ageing

conditions.

5.2.2 Simulation results

Figure 5.10 shows the results of the computational analysis of the processing

route. Figure 5.10a is the temperature profile of the aTMP route, which is the same

as used during the experimental study. This includes casting, solution annealing,

water quenching, reheating to deformation temperature and soaking prior to

deformation, deformation, direct ageing, and air cooling. The additional cooling

and heating cycle after casting and before solution annealing is skipped to reduce

computing time. This is based on the assumption that precipitates will already form

during solidification, and not be dissolved during further cooling and heating. The

evolution of the precipitate mean radius, number density and phase fraction over

time can be seen in Figure 5.10b-d. From these graphs, it can be observed that

TiN precipitates already form during cooling from solution annealing, which is


76

denoted as TiN(gamma) in the simulation, to emphasize precipitation in the

austenite. These precipitates will not be dissolved in the following aTMP. During

reheating after quenching, cementite (cem) starts to precipitate and grow during

soaking time prior to deformation. It can be observed that NbC starts to precipitate

already during the cooling from solution annealing temperature highlighted by the

increase in number density (Figure 5.10c) too. The uneven line in the mean radius

plot (Figure 5.10b) for the NbC precipitates is due to multiple nucleation

avalanches, which lead to multiple decreases of the mean radius due to the

accumulation of newly formed small particles. A considerable jump in number

density, phase fraction and mean radius of both, NbC and TiN(alpha), denoting

precipitation in the ferrite, can be observed during deformation due to strain

induced precipitation (highlighted with orange circles in Figure 5.10b-d). The final

mean radius of the cementite after 60 min ageing is the largest with 104 nm,

followed by the TiN(gamma) that presumably formed in the austenite. TiN(alpha)

and NbC remain rather small with an average final mean radius of around 2 nm. It

should be noted that the Ti- and Nb-rich precipitates show a core-shell structure

in the experiment. In the simulation, this fact cannot be accounted for since

MatCalc does not provide a corresponding precipitation model. Since the NbC

precipitates are treated as separate precipitate populations, and many small

precipitates nucleate in a later stage of the heat treatment, the prediction for the

number-weighted mean radius delivers too small radii compared to the

experiment.

The evolution of the subgrain diameter over time is depicted in Figure 5.10e.

Starting with a large starting subgrain diameter of ~100 µm it can be seen that the

onset of subgrain formation correlates with the deformation segment. The

dislocations introduced during deformation enable subgrain formation. The

simulation shows that this microstructure remains stable upon ageing and only a

minimal increase in subgrain diameter was seen. The final subgrain diameter

simulated after 60 min ageing was 0.84 μm.


77

Figure 5.10 - MatCalc-Simulation of aTMP with 60 min direct ageing time, (a) temperature

profile of aTMP route, (b) evolution of mean radius of precipitates over time, (c) evolution

of precipitate number density over time, (d) evolution of precipitate phase fraction over

time, (e) evolution of subgrain diameter over time.


78

Table 5.1 shows the mean chemical composition of the precipitates simulated after

60 min ageing time. It can be seen that the cementite is enriched in Mn. In the

NbC a considerable amount of Ti is observed with 10.5 wt.%.

Table 5.1 - Mean chemical composition in wt.% after 60 min direct ageing.

Fe C Mn Nb Ti N

cementite 62.5 6.61 26.72 4.13 0 7.97 e-5

TiN

(gamma)

5.54 e-6 0.8 8.92 e-6 0.56 76.97 21.67

TiN

(alpha)

1.81 e-10 0.31 7.53 e-10 7.9 e-3 77.4 22.3

NbC 1.69 e-6 11.69 2.15 e-6 76.82 10.5 0.97

Table 5.2 shows the simulated evolution of precipitate diameter with increase in

direct ageing time. Obviously, all precipitates undergo coarsening during the direct

ageing, but this is most significant for cementite starting at 64 nm in as-deformed

condition and reaching an average size of 104 nm after 60 min ageing. TiN is much

larger with about 42 nm compared to the TiN precipitated in the ferrite with 1.9 nm

after 60 min ageing.

Table 5.2 - Simulation results for precipitate diameter after different ageing times.

Precipitate

diameter in nm

0 min 10 min 60 min

cem 64 96 104

TiN(gamma) 40 40 42

TiN(alpha) 1.4 1.8 1.9

NbC 1.4 1.9 1.9


79

5.3 Discussion

The overall goal which has been achieved in this chapter was to obtain a

homogeneous hierarchical microstructure in an HSLA steel via warm deformation

in the ferrite. The individual structural components will be discussed in the

following section and compared to MatCalc modelling results.

5.3.1 Grain size and texture

An advanced thermo-mechanical processing route was designed to obtain a

homogenous microstructure in the deformation zone of a Ti-Nb-Mo HSLA steel

during plane strain compression tests in a Gleeble thermo-mechanical simulator

(see Figure 5.1c-f). The successful processing route, see Figure 3.3, consists of

10 min soaking time prior to warm deformation at 600°C followed by three

deformation passes with a strain of 0.2 each, 30 s interpass time and a strain rate

of 1 s- 1. Compared to our previous study [90], several changes to the processing

route were made in order to achieve more homogenous deformation. To avoid the

formation of a macro shear band, as observed previously during single pass

deformation, the maximum strain applied per pass was kept at 0.2 and the strain

rate was reduced from 10 s-1 to 1 s-1. The rationale behind these changes is that

in our previous study shear band formation initiated at a true strain between 0.2

and 0.3. Furthermore, Xu et al.[85] reported shear band formation in a low carbon

steel starting at high strain rates and strains exceeding 0.3. In our study, the

interpass time was increased to 30 s, in order to enable limited recovery and/or

recrystallisation [91,92]. Furthermore, a longer soaking time of 10 min before

deformation was found to additionally decrease the tendency for shear band

formation, due to softening during tempering the martensitic/bainitic starting

microstructure. Tempering of martensite leads to increased toughness due to

precipitation of cementite particles and recovery/recrystallisation of the

martensite [66,93]. Although only an overall true strain of 0.6 was applied,

microstructures with grain sizes of around 0.55 μm were achieved and appear to

be stable upon ageing (see Figure 5.2a), which is similar to our previous study

with typical true strains of 0.8. Other authors achieved similar or larger grain sizes

between 0.3-1 µm in microalloyed steels for various other approaches of aTMP,

such as cold rolling and annealing as well as warm deformation in the austenite-

ferrite two phase region [6,8,52,54,55,69]. However, from the EBSD scan shown


80

in Figure 5.1f, it can be seen that microstructure mainly consists of subgrains that

are confined mostly by LAGBs. Only a few fully recrystallised grains confined by

HAGBs arranged in necklace-like structures can be found near the original high

angle grain boundaries and regions with local C enrichment. A local change in

recrystallisation mechanism to discontinuous dynamic recrystallisation is assumed

to be present here, however, this is beyond the scope of the present study and will

be discussed in more detail in a future study. However, a possible reason for the

high fraction of LAGBS in the surrounding microstructure may be that the applied

total true strain of 0.6 is not sufficient for continuous dynamic recrystallisation to

reach its steady state with a 1 : 1 ratio of HAGBs : LAGBs [68]. The experimental

grain size is slightly smaller compared to the simulation results which predict a

subgrain diameter of ~0.8 μm (see Figure 5.10). This may be likely because the

simulation does not take into account the conversion from LAGBs to HAGBs

during continuous dynamic recrystallisation. Further, the resulting grain size

determined by EBSD is largely dependent on the detailed data evaluation

procedure.

Due to the limited sample volume probed by EBSD in our study, we suggest using

the resulting texture mainly to confirm the success of aTMP design rather than for

extended discussions of the detailed textural evolution in this particular steel. The

rationale behind this is that localised deformation is likely to result in a more

random texture than homogeneous deformation, which will result in a typical bcc

rolling texture. To explain typical bcc rolling textures, Rosenberg et al. [94]

calculated the Taylor factor and lattice rotations for bcc metals deformed by pencil

glide. They showed that <110>//ND is not stable during deformation and,

therefore, these grains will rotate towards <111>//ND or <100>//ND. <111>//ND

grains are known to have a high Taylor factor and require activation of many glide

systems to undergo further deformation. Therefore, they are considered as hard

and high stored energy regions. In contrast, <001>//ND grains have a low Taylor

factor and are considered as soft, low stored energy regions. This behaviour was

also observed in our experiments as can be seen in Figure 5.1f and Figure 5.3.

However, two of our samples do not show such a typical texture formation as

exemplarily shown in Figure 5.3c. Possible explanations are the nature of the

plane strain compression test in the Gleeble, with its limited deformation zone and

special tool geometry, resulting in a dog-bone shaped cross-section of the


81

as-deformed materials, as modelled by Aksenov et al. [84]. Here, the material

volume close to the edges of the tool is prone to undergo localised deformation.

Further, this could be an effect of the limited scan area picked up by EBSD. Larger

area EBSD scans should be carried out on up-scaled samples processed by rolling

but this is beyond the scope of this study.

Overall, there are several studies on the competition of <001>//ND versus

<111>//ND grains at deformation temperatures in the ferrite region, and most of

these authors suggest a tendency towards formation of a strong <111>//ND

texture.

Anijdan et al. [95] studied the influence of a small true strain of 0.2 applied at 400°C

on three different microalloyed steels and reported that the texture that developed

was preferably <111>//ND texture. Han et al. [96] investigated the changes of

texture during annealing at 650°C of a Ti-containing interstitial-free steel after

warm rolling at a finish rolling temperature of 580°C with 90% reduction in

thickness. They found that a typical bcc rolling texture was achieved. <111>//ND

orientated grains preferably underwent recrystallisation. With increasing ageing

time, the amount of <001>//ND oriented grains were consumed by <111>//ND

recrystallised grains and, therefore, a more pronounced γ-fibre was formed.

Toroghinejad et al. [97] studied the influence of warm rolling between

temperatures of 440-850°C with a reduction of 65% in thickness for different

interstitial-free and low carbon steels. They reported a favourable <111>//ND

texture. Furthermore, NbC precipitates retarded recrystallisation, leading to

increased strain accumulation and, thus, to more shear bands as nucleation sites

for <111>//ND grains. This has been also reported by Barnett et al. [98] and

Humphreys et al. [99] who suggest that in-grain shear bands provide more

nucleation sites for developing γ-fibre texture and, therefore, increased deep-

drawability.

Overall, in our samples, we did not observe a strong trend to a more pronounced

<111>//ND texture. One reason could be that we did not observe in-grain shear

band formation although fine precipitates are present in the microstructure, but

they may not yet be fully precipitated when the materials undergo dynamic

restoration processes, as evidenced by the MatCalc simulation (Figure 5.10).


82

5.3.2 Precipitates

Two types of precipitates were found: Larger (>50 nm) round particles which

preferably nucleate at grain boundaries and smaller (<50 nm) cuboidal

precipitates, which nucleate within the grains on dislocations and subgrain

boundaries.

5.3.2.1 Large (>50 nm) precipitates:

In the TEM bright and dark field images in Figure 5.4a-b and Figure 5.5a-b, the

precipitates at the grain boundaries are large and of round/elliptical shape. The

TEM EDS mapping in Figure 5.6a as well as the selected area diffraction in Figure

5.7 reveal that the larger, round precipitates are cementite. The EDS mapping

shows that these types of precipitates are rich in Fe, Mn and C whereas no other

elements were detected. This matches well with the simulation results concerning

the chemical composition of the precipitates (see Table 5.1) showing that

cementite incorporates a significant amount of Mn. Furthermore, the simulated

results show that the mean radius of cementite is significantly larger with 104 nm

compared to TiN and NbC with 2 nm after 60 min ageing (Figure 5.10b and Table

5.2) which was also observed in the experimental study. The computed result of

104 nm after 60 min ageing for cementite matches quite well with the second peak

in the experimentally measured precipitates size distribution in Figure 5.9 at

around ~120 nm. However, one needs to be aware that this peak is quite broad

with the largest measured precipitate being 630 nm. Furthermore, due to the rather

complex precipitation sequence, it is not possible to perfectly distinguish between

the two types of precipitates experimentally, as EDS analyses of each individual

particle would be required.

Interestingly, a large quantity of cementite in close proximity, as seen in Figure

5.4a-b, is observed in TEM. This might be because of C segregation making it

more easily available to form cementite in this region. Another explanation could

be that the martensitic/bainitic starting microstructure underwent tempering before

deformation [66,93] leading to these areas with high cementite density.

5.3.2.2 Small (<50 nm) precipitates:

The TEM EDS mapping in Figure 5.6 shows that the smaller cuboidal precipitates

are rich in Ti and Nb. The simulation also suggests that TiN and NbC precipitates


83

form during this processing route. The computed mean radii of TiN and NbC are

significantly smaller than cementite (see Table 5.2 and Figure 5.10b), which

matches well with the experimental observations. It is well known that Ti forms TiN

at rather high temperatures (<1500°C), which is favourable for pinning of austenite

grain boundaries if deformation occurs in the austenite [16]. This is also evidenced

by thermo-kinetic modelling (see Figure 5.10). However, if the material undergoes

deformation in the ferrite region, the previously formed TiN particles will mainly act

as nucleation sites for NbC resulting in large, complex precipitates [100,101]. This

can be seen in Figure 5.5 and Figure 5.6, where NbC segregated around a square-

shaped TiN to form a star-shape. Similarly, Jia et al. [100] studied the precipitation

behaviour of two Ti-Nb microalloyed steels. They reported that NbC precipitates

epitaxially grew on the surface of a TiC due to the small lattice mismatch.

Kapoor et al. [101] investigated the chemical composition of two Nb-Ti

microalloyed steels using atom probe tomography. They reported that carbo-

nitride particles that are rich in Ti and N but lean in C precipitate at high

temperatures. With decreasing temperatures, shells of (Ti)(C,N) will form on the

already existing precipitates. These layers will become richer in C and leaner in N

with decreasing temperature, and eventually Nb starts to segregate on the outer

layers. They also suggest that this mechanism will deplete Nb from the

surrounding matrix, making it less favourable for low-temperature

precipitation [101]. The computed size of the TiN(gamma) is 40 nm in the

as-deformed condition and increases slightly to 42 nm after 60 min ageing. This

matches well with the first peak of the experimentally measured precipitate size

distribution seen in Figure 5.9. TiN and NbC are computed as separate

precipitates, therefore, the core-shell formation as seen in Figure 5.5c-d and

Figure 5.6 cannot be predicted. This explains why the precipitate size of NbC and

TiN(alpha) (see Table 5.2) is underestimated by the simulation. Another reason

might be the presence of small clusters which are beyond the detection limit of the

TEM used here. Such features could be detected by three-dimensional atom

probe microscopy, but this is beyond the scope of the current study.

Thus, for future alloy development of HSLA steels with martensitic/bainitic starting

microstructure for aTMP in the ferrite region, it is suggested to replace Ti by other

micro-alloying elements. As deformation does not occur in the austenite, the main

purpose of TiN, which is stabilizing the austenite to prevent grain growth, is not


84

required in this case. In the ferrite, TiN will mainly undergo coarsening and remove

Nb from the matrix, hence, making Nb unavailable for further precipitation of

smaller particles.

To conclude, we demonstrated a novel processing route that leads to a

homogenous hierarchical microstructure without the formation of macroscopic

shear bands. Furthermore, successful through-process thermo-kinetic modelling

was performed using the software MatCalc which will be able to guide future alloy

design work. Furthermore, a deeper understanding of the detailed precipitation

sequence has been established. A demerit was the composition of the alloy used

here as TiN present in the current steel favoured the formation of Nb-rich shells

around these pre-existing particles. Therefore, future research in this area where

deformation occurs mostly in the ferrite will focus on alloy design of modern HSLA

steels without Ti. In a next step, this aTMP will be up-scaled to a rolling mill to

enable a thorough study of the mechanical properties.


In the present chapter, an optimised aTMP route was developed that results in a

homogenous hierarchical microstructure in a modern Ti-Mo-Nb HSLA steel. The

investigation is based on a thorough experimental and modelling approach

delivering:

A soaking time of 10 min at 600°C prior to warm deformation, as well as a

three-pass deformation pass with a strain of 0.2 per pass with a strain rate

of 1 s-1 allows to achieve a homogenous deformation during plane strain

compression testing.

Continuous dynamic recrystallisation is the main dynamic restoration

mechanism active which lead to grain refinement with (sub)grain sizes

around 0.55 µm according to electron microscopy and 0.8 µm according to

modelling with MatCalc.

Two types of precipitates are observed: Larger (>50 nm) round FeMnC-rich

precipitates, which preferably nucleate at high angle grain boundaries, and

smaller (<50 nm) cuboidal TiNb-rich precipitates also nucleating at

dislocations and subgrain boundaries. It is shown that the TiNb particles

undergo a shape and composition change forming a core-shell structure


85

when subjected to direct ageing after deformation. This is attributed to

diffusion of C, Si, N, and Nb and relatively higher stability of NbC at lower

temperatures.

Thermo-kinetic modelling using MatCalc is shown to enable a detailed

through-process study of the microstructural evolution during aTMP route.

Results match well with the experimental data.

In the following chapter, the aTMP processing route developed in this chapter

using the Gleeble thermo-mechanical processing simulator is translated to a

larger-scale rolling mill. This enables the study of the industrial feasibility and

thorough testing of mechanical properties (tensile and Charpy impact testing) on

large samples.

Initial report on structure-property relationships of HSLA steels subjected to aTMP ________________________________________________________________________________________________________________________________

86

6 Initial report on structure-property relationships of

HSLA steels subjected to aTMP


C. Ledermueller, H. Zhu, H. Li, S. Primig, An Initial Report on the Structure–

Property Relationships of a High-Strength Low-Alloy Steel Subjected to Advanced

Thermomechanical Processing in Ferrite, Steel Res. Int. 1900596 (2020)

C. Ledermueller did the majority of the planning and execution of the

experimental work as well as the data analysis. The manuscript was drafted by

her.

6.1 Introduction

In the previous chapter, it was shown that through an optimised aTMP design a

homogenous deformation microstructure of a HSLA steel can be achieved during

plane strain compression testing. Continuous dynamic recrystallisation is the

primary mechanism responsible for the grain refinement to (sub)grain sizes of

around 0.55 µm. Two types of precipitates were found: larger (>50 nm) round

FeMnC-rich and smaller (<50 nm) cuboidal TiNb-rich precipitates, which also

correlates with the results in Chapter 4. Additionally, it was found that the TiNb

particles undergo a change in shape and morphology exhibiting a core-shell

structure when subjected to direct ageing at 600°C. Furthermore, it was shown

that the microstructural evolution can be modelled in detail using the thermo-

kinetic modelling software MatCalc, in good agreement with the experimental data.

The aim of the current chapter is to up-scale this processing route to carry out

thorough mechanical testing in order to verify the success of the process design

regarding strength, ductility and target microstructures. Using the Gleeble in plane

strain compression mode has its limitations in simulating the strain distribution

obtained under real rolling conditions. Therefore, it is necessary to compare the

microstructures and textures achieved during plane strain compression

(Chapter 5) and rolling.


87

Furthermore, typical sample sizes for thermo-mechanical processing simulators

such as a Gleeble are small. Hence, most of the mechanical testing is restricted

to hardness testing or using sub-size sample geometries for tensile and impact

toughness testing. Qun et al. [50] used dog-bone shaped tensile samples with

geometries of 6 mm length, 2 mm width and 1 mm thickness. Song et al. [9] used

Charpy sub-size specimens with a cross-section of 3 x 4 mm2. Sub-size

specimens might suffer from size effects leading to over/under-estimation the

actual large-scale mechanical properties of the steel [102–104]. Therefore, there

is a need for mechanical data on warm-deformed low-alloyed steels which are

obtained from standardised tests (see Chapter 3.6). In this chapter, it is indented

to overcome these challenges by up-scaling advanced thermo-mechanical

processing using a Hille 100 rolling mill. The processing applied in this chapter is

described in Chapter 3.3.3.

Mechanical properties of advanced thermo-mechanical processed low-alloyed

steel can vary, depending on the exact processing and steel chemistry and have

been reported in the range of 480-898 MPa for the ultimate tensile strength (UTS)

with total elongation of 11.6-27%, [8,9,49,50,105]. It has been shown that cold

rolling and annealing of a martensite [58,63,64] or tempforming of high C steels

via large strain calibre rolling [45] can lead to exceptional strengths of up to

1.5 GPa. In ultrafine-grained steels, the upper shelf energy is relatively low due to

delamination [4,9,70]. However, delamination has been shown to be beneficial for

low temperature toughening and decreasing the ductile-to-brittle transition

temperature under certain circumstances. This has been referred to as

delamination toughening [71,106]. Hence, in this chapter, it will be evaluated if

delamination toughening occurs under warm-rolling conditions without heavy

deformation (e.g. 55% reduction of thickness as compared to <80%) for the same

steel as used in the previous chapters.

6.2 Results

6.2.1 Microstructure

The inverse pole figure maps depicted in Figure 6.1 show the microstructures of

the as-rolled and 60 min aged conditions. Typically, the grains are elongated in

rolling direction with the rolling texture that is commonly observed for bcc metals.


88

The majority of the microstructure is banded due to elongated grains with some

areas being more significantly recovered. In certain areas, new grains confined by

HAGBs have formed. Some smaller recrystallised grains (white arrows) can be

seen along the original grain boundaries in both conditions. The grain size in as-

deformed condition is 0.77 µm and increases to 1 µm after 60 min ageing. It is to

mention that this grain size includes all crystallites that are confined either by

LAGBs or HAGBs and it is an average of the small recrystallised grains and the

large original grains. However, upon ageing, some of these recrystallised grains

(black arrow) grow significantly and can reach diameters around 4-5 µm after

60 min at 650°C. The red arrow in Figure 6.1b points out a recrystallised grain that

formed inside an original grain. The STEM images and the EDS maps in Figure

6.2 show examples of the two distinct types of precipitate, which have also been

observed previously, larger (>50 nm) FeMNC-rich precipitates along grains

boundaries and some smaller (<50 nm) TiNbC-rich precipitates inside the grains

in the as-rolled sample. The diameter of 567 precipitates for the 0 min ageing

condition and 354 precipitates for the 60 min ageing condition were measured

from STEM images. The average precipitate size for the 0 min condition is

51±38 nm with a minimum size of 4 nm and a maximum size of 251 nm. For the

60 min sample, the average precipitate size is 84±70 nm with a minimum size of

6 nm and a maximum size of 474 nm.


89

Figure 6.1 - Exemplary EBSD IPF maps of the a) as-rolled and b) 60 min directly aged

samples. High angle grain boundaries are shown as black lines, low angle grain

boundaries are white lines. The inset shows IPF colouring code in normal direction

(vertical). The white arrows show small recrystallised grains along the original grain

boundaries, the black arrow shows larger recrystallised grains and the red arrow shows a

new grain that formed due to static Rxx.


90

Figure 6.2 - BF-STEM images for the as-rolled condition a)-b) larger FeMn-rich

precipitates at grain boundaries, c)-d) smaller TiNb-rich precipitates inside grains, e) ADF-

STEM and EDS elemental maps showing both FeMn-rich and TiNb-rich precipitates.


91

6.2.2 Mechanical Properties

Figure 6.3 depicts typical engineering stress-strain curves for all 4 ageing

conditions with the values for lower yield strength, ultimate tensile strength (UTS)

and total elongation summarised in Table 6.1. It can be seen that the lower yield

strength and the UTS for the as-rolled samples in the highest with 622±3 MPa and

654±5 MPa, respectively. The UTS drops by 7.6% to 604±3 MPa after 60 min

ageing. The total elongation is the lowest for the as-rolled sample compared to the

other ageing conditions with 14±0.5 %. After 60 min of direct ageing, the total

elongation reaches 16±0.4 %. However, all of these steels exhibit a high yield ratio

with around 0.95.

Tensile testing:

Figure 6.3 - Exemplary engineering stress-strain curves for HSLA steels subjected to

different direct ageing, 10 min is dashed because sample started necking outside the

gauge length.


92

Table 6.1 - Results of tensile testing showing lower yield strength, ultimate tensile strength

(UTS) and the total elongation for the as-rolled condition and all direct aged samples. The

value for the 10 min sample is greyed out as the necking occurred outside the gauge

length.

lower yield strength

[MPa]

UTS [MPa] total elongation

[%]

0 min 622±3 654±5 14±0.5

10 min 598 639±0.7 no value

30 min 588±0.7 624±4 15±0.1

60 min no value 604±3 16±0.4

Figure 6.4 shows exemplary macroscopic and SEM images of the fracture

surfaces after tensile testing of the as-rolled and 60 min aged samples.

Delamination (black arrows) can be observed across the width of both sample

conditions as seen in Figure 6.4a and c. Furthermore, both fracture surfaces

consist of a mixture of cleavage fracture (blue arrow) and dimple formation (red

arrow) in Figure 6.4b and d.


93

Figure 6.4 - Typical tensile testing fracture surfaces for a)-b) as-rolled condition and c)-d)

60 min direct ageing condition. Images a) and c) were obtained using photography and b)

and d) are secondary electron images obtained in the SEM. The rolling direction is out of

the plane/ towards the reader. Black arrows show delamination, blue arrows show

cleavage fracture and red arrows show dimple formation.

Charpy impact testing:

Figure 6.5a-b shows the evolution of impact energy over temperature for subsize

specimens in longitudinal and transversal directions with regard to the rolling

direction. In Figure 6.5a it can be seen that the longitudinal samples show a higher

impact energy as compared to the transversal samples. Figure 6.5b shows that at

-90°C the impact energy is similar for all samples in longitudinal direction but that

it increases with increasing temperature. Overall, it can be seen that the 30 min

and 60 min ageing samples have the highest impact energies at room temperature

with 114±33 J and 197±6 J respectively. The ductile to brittle transition

temperature (DBTT) is the temperature at half of the upper shelf energy. The

DBTT for the 60 min and 30 min sample is around -22°C. The fracture surfaces

for the 30 min and 60 min samples at -50°C are shown in Figure 6.5c-d. They


94

appear predominately brittle and, therefore, support this result. However, please

keep in mind that this is a rough estimation as more testing in the transition region

would be needed.

Figure 6.5 - Impact energy over temperature for subsize Charpy V-notch samples: a) in

longitudinal and b) longitudinal and transversal direction for the 0 min and 60 min

conditions, c) fracture surface for 30 min longitudinal direction at -50°C, d) fracture surface

for 60 min longitudinal direction at -50°C.

Figure 6.6 shows macroscopic as well as SEM images of the fracture surfaces of

the as-rolled and 60 min ageing samples at 3 different temperatures, -90°C, room

temperature (RT) and 120°C. The -90°C samples predominantly exhibit cleavage

fracture whereas with an increase in temperature a higher amount of ductile

fracture was observed, indicated by an increased amount of dimple formation. The

biggest difference in the morphology of the fracture surfaces can be seen at room

temperature. The 0 min sample shows delamination whereas the 60 min aged

sample does not. This corresponds well with the higher impact energies measured

for this condition at room temperature.


95


96

Figure 6.6 - Typical fracture surfaces after impact testing at different temperatures (see

insets) for the longitudinal directions of a) as-rolled and b) 60 min ageing conditions.

Images were obtained by photography and by SE-SEM. Red arrows show dimple

formation and blue arrows show cleavage fracture.

From Figure 6.7, it can be seen that delamination is independent of sample

direction and occurs in similar manner in the longitudinal and transversal samples,

not necessarily parallel to the rolling direction.


97

Figure 6.7 - Delamination for the 30 min ageing sample at room temperature in a)

longitudinal and b) transversal directions.

6.3 Discussion

Microstructure:

The microstructure obtained after rolling (see Figure 6.1) is significantly different

to the microstructure obtained after thermo-mechanical processing in a Gleeble

(Chapter 5) [107]. This suggests that the applied strain during rolling is not

sufficient to provide fine grains across the whole sample as it was suggested after

the Gleeble experiments in Chapter 5. This also highlights the limitation of Gleeble

testing which shows flow inhomogeneities [84] as compared to rolling. In a Gleeble

the strain in plane strain compression mode is more localised and therefore

probably higher as applied and leads to more uniformly UFG. However, no macro

shear banding was overserved during rolling. Overall, the recrystallisation

mechanism present in this study is CDRX which has not yet reach steady

state [68]. Additional discontinuous dynamic recrystallization (DDRX) can be seen

due to the formation of newly grains confined my HAGBs along original HAGBs

(necklace structures), around in-grain shear bands/microbands and in areas of

local C segregation [91,108–110]. The newly formed grain highlighted with a red

arrow in Figure 6.1b most likely formed due to static Rxx.

Tensile Testing:

The obtained UTS was 650 MPa, the elongation 14% and the yield ratio 0.95 in

as-rolled condition (Figure 6.3 and Table 6.1). This UTS would correspond to a

Vickers hardness of ~202 HV10 according to ISO 18265 table A.1. This is

significantly lower as compared to the 965 MPa in Chapter 4. This higher value is

likely due to the strain localisation in the shear band and the initial higher applied


98

strain of 0.8 in Chapter 4. This is compared to results achieved for similar

processed low-alloyed steels in the following:

Qun et al. [50] achieved a slightly lower UTS of 570 MPa and elongation of 16%

with deformation at 600°C, strain of 0.7 and strain rate of 0.001s-1 for a

0.17C-0.37Si- 0.68Mn-0.036P-0.039S (wt%) steel using a Gleeble. The lower UTS

by Qun et al. [50] could be explained by the slightly lower strain applied and the

lack of microalloying elements such as Nb and Ti as compared to our study.

Calcagnotto et al. [49] applied overall strain of 1.6 at 550°C and achieved a UTS

of 633 MPa and elongation of 13.3% with a yield ratio of 0.91 for a 0.17 C-1.49

Mn-0.22 Si-0.033 Al-0.0033 N-0.0017 P-0.0031S (wt%) dual-phase steel, which is

similar to our results. However, they applied twice as much strain as compared to

our study. Cheng et al [8] achieved a yield ratio of 0.95 and an elongation of 14.7%

in a Fe-0.11C-0.21Si-1.48Mn-0.11Ti-0.31Mo-0.041Al-0.0028S-0.0054P steel

after a strain of 55.6% at 850°C and a strain of 65% at 650°C followed by 30 min

ageing at 600°C. However, they obtained a UTS of 898 MPa, which is 250 MPa

higher as compared to our study. Cheng et al. [8] report that 'superdense'

microbands, are contributing to strengthening, besides precipitation enhanced

grain refinement. They suggest that the formation of these microbands can be

suppressed by stress fields around precipitates which promote the formation of

HAGBs, especially in the γ–fibre [8]. Further, they observed in another study that

after an increase in warm rolling temperature to 700°C, superdense microbands

did not form [111]. According to Humphreys and Hatherly [112] the misorientation

of microband walls does not significantly change with strain whereas in microshear

bands there is an increase in misorientation with an increase in strain. This implies

that at moderate strains microbands are transient microstructural features and

only become persistent above strains of 1.5. Microshear bands however are

permanent features within the microstructure although their shear and

misorientation increases with strain [112]. Therefore, it would be necessary to

study the shear deformation in our material with varying strains in order to be able

to identify them as microbands or microshear bands. Song et al. [9] achieved a

total elongation of 20% and UTS of ~600 MPa for a Fe-0.22C-0.21Si-0.74Mn-

0.004P-0.003S-0.001N-0.029Al (mass%) subjected to warm deformation at 550°C

with a total strain of 1.6 followed by annealing for 2h. This shows a slightly lower

UTS but larger elongation as compared to our study. The lower UTS can be


99

explained by the absence of microalloying elements in their steel. The total

elongation is around 6% larger as compared to our study, due to the higher work

hardening rate reported in Song’s steel. The low yield-ratio of 0.95 (Figure 6.3)

achieved in our study can be explained by the rather large precipitate size and low

number density (see Figure 6.2). Thus, precipitates are not significantly contribute

to hardening. We suggest that for further research it would be more beneficial to

have an alloy concept without Ti, as TiN is formed at higher temperatures which

cannot be dissolved during further processing. Further, in a previous study by

some of the current authors [107], it was observed that Nb segregates to these

already existing TiN, which reduces its availability to form fine nanoscale

precipitates. Ghosh et al. [105] studied the mechanical properties of a Ti-Nb

stabilized IF steel warm rolled at 650°C with a reduction in thickness of 50% and

80%. For the 80% reduction, they achieved a UTS of 479±3 MPa, a yield strength

of 421±5 MPa and an elongation 27±2%. The ~150 MPa lower UTS as compared

to our study can be explained by the lack of C to form precipitates which would

contribute to strengthening. The exceptional elongation which is almost twice as

much as we achieved can be explained due to bimodal grain sizes with ultrafine

grains around 1-3 µm and large grains 25-27 µm in their steel [105]. They assume

that the smaller grains contribute to the increase in yield strength whereas the

larger grains provide ductility.

Impact toughness:

Due to the small sample volume usually obtained when studying aTMP at labscale

not many reports on the impact toughness of advanced thermo-mechanically

processed low-alloyed steels are available.

At room temperature, the impact toughness of the sample aged for 60 min is

significantly higher as compared to the other sample conditions due to the absence

of delamination (Figure 6.5 and Figure 6.6b). Das et al. [113] found that crack

propagation in an (oxide dispersion strengthened) ODS steel with bimodal grain

size distribution usually follows the UFGs as they do have a lower deformation

capability, hence initiating void formation. Therefore it is suggested that the

reduced amount of small grains and the presence of the larger recrystallised grains

might be responsible for the higher impact toughness at room temperature.


100

Song et al. [9] found the DBTT in sub-size samples of a Fe-0.22C-0.21Si-0.74Mn-

0.004P-0.003S-0.001N-0.029Al (mass%) with a grain size of 1.3 µm to be

at -120°C. This is around 100°C lower than in our study (see Figure 6.5). They

suggest that delamination in their samples are beneficial in lowering the DBTT.

The detailed mechanism for that is yet unclear but might be due to their distorted

ferrite-pearlite microstructures, elongated ferrite grains, texture effects or

particles [4]. In our study, we found a similar delamination behaviour although

independent of grain direction (transversal vs longitudinal) as can be seen in

Figure 6.7. One reason might be that there a different types of delamination

present such as the rack arrester or crack divider type although in the transversal

test direction used in our study the weak planes are parallel to the testing direction.

Therefore, we suggest that the presence of particles or the texture are more likely

to be responsible for the delamination. However, in our study, it is unclear whether

delamination is beneficial in lowering the DBTT as with just -22°C it remains fairly

high.

The inverse temperature dependence of the impact toughness as suggested by

Kimura et al. [114] and Min et al. [75] was not observed in the current study, mainly

because we did not observe a strong <110> // RD fibre texture but a mixture of

<110> // RD and <100> // RD textures (see Figure 6.1).

Wang et al [115] studied the impact toughness of two Mo-containing low-alloyed

steel with one of them additionally containing Nb which were tempered at 700°C.

They found that the combination of Nb and Mo can lead to an increased impact

toughness as Mo-rich NbC impeding Mo segregation to grain boundaries where it

would precipitate as ξ-particles. However, further studies would be required to

prove this for the material used in our study.


Overall, these results are comparable with what has been reported in literature,

however, the difference is that only a total reduction in thickness of around 55%

which equals to a true strain of 0.8 was used in this study. This is roughly half of

the strain other research groups used in their studies, thus, highlighting that the

approach studied in Chapter 4-6 is indeed very promising.


101

It was possible to successfully upscale the aTMP route with no strain

localisation using a Hille 100 rolling mill.

A UTS of 650 MPa with a yield ratio of 0.95 and an elongation of 14% was

obtained in as-rolled condition.

The DBTT achieved is relatively high with -22°C, hence it is suggested that

delamination did not contribute to toughening.

After 60 min ageing the impact toughness at room temperature is

significantly improved due to the formation of large (~4 µm) recrystallised

grains.

Delamination did occur similar in transversal and longitudinal directions,

therefore we suggest that grain orientation is not responsible for this

phenomenon.

For future work, it is suggested to more carefully study the impact toughness in

the area of the DBTT and compare the results with a coarse-grained steel. Further

it is suggested to focus on a different steel composition without Ti addition to

guarantee a smaller precipitate size and hence have a higher work hardening rate.

Finally, a slight increase in applied strain might be beneficial to obtain more

recrystallised grains and hence a lower DBTT.

Influence of Cr/Mo as microalloying elements in HSLA steels subjected to warm deformation ________________________________________________________________________________________________________________________________

102

7 Influence of Cr/Mo as microalloying elements in

HSLA steels subjected to warm deformation


C. Ledermueller, H.I. Pratiwi, R.F. Webster, M. Eizadjou, S.P. Ringer, S. Primig,

Microalloying effects of Mo versus Cr in HSLA steels with ultrafine-grained ferrite

microstructures, Mater. Des. 185 (2020).

C. Ledermueller did a major part of the experimental work as well as the data

analysis. The manuscript was drafted by her. She supervised the work done by

H.J. Pratiwi.

7.1 Introduction

In the previous chapters, it was shown the industrial Ti-Nb-Mo HSLA steel with the

composition shown in Table 3.1 is suitable for aTMP schedules (see Chapter 5)

[90,107]. Hierarchical microstructures, comprising grain and subgrain boundaries

decorated by cementite and complex core-shell (Ti/Nb)(C,N) particles formed

during ageing were observed. Interestingly, fine-scale precipitates that might have

formed in the ferrite were not observed. This suggests that Nb may be a more

promising candidate than Ti for HSLA steel design via aTMP as Nb will remain in

solid solution and therefore available for strain-induced precipitation during

ageing.

The key concept of the current chapter was to extend the above approach to

explore the potential for additions of so-called ‘modern’ microalloying elements

such as Mo and Cr to provide further strengthening [18]. The background to our

compositional design was as follows: Firstly, Nb is one of the most common and

well-studied microalloying elements in HSLA steels. The main benefit of Nb is its

strong tendency to form Nb(C,N), which does not only contribute to precipitation

strengthening, but also pins grain boundaries enabling finer grain sizes and

delaying the γ-α phase transformation. Furthermore, the solute drag effect of Nb,

where solute Nb segregates to dislocations and grain boundaries, is a major

contributor to increased strength and delayed recrystallisation [13–16]. The


103

addition of Mo as a microalloying element is known to provide solid solution

hardening, increase hardenability, and to delay both phase transformations and

recrystallisation [17,18]. However, one of the most beneficial effects of Mo

microalloying was found in conjunction with Nb, modifying the precipitation

sequence [18] and leading to an increase in yield strength. The detailed role of Mo

in the precipitation process is not understood, but its effect is to clearly refine the

precipitate dispersion upon ageing. There are reports that Mo (i) reduces the

diffusivity of other carbide forming elements [17,24], (ii) segregates to the

precipitate/matrix interface between precipitate and matrix [25], (iii) is incorporated

into the NbC composition [26–28], and (iv) acts as a nucleation site for

NbC [18,25]. Most studies investigating the effects of Mo were conducted on TiMo

microalloyed steels [27,29,30]. Park et al. [31] studied a 0.4Nb-0.2Mo

microalloyed steel which was hot rolled and then aged at temperatures between

500-700°C. They did not observe Mo within NbC precipitates but rather found it in

solid solution in the matrix. The addition of Cr was of interest to our program due

to its ability to enhance corrosion resistance [16]. It is also well-known that

microalloying of Cr in low C steels can promote the formation of Cr-carbides which

contribute to strengthening [33–36]. Moreover, Cr can inhibit cementite coarsening

due to the dual partitioning behaviour of Cr and Mn between cementite and the

ferrite matrix [19,20].

This chapter studies the relative effects of microalloying with Mo versus Cr on the

microstructural evolution and properties of three model Fe-1.6Mn-0.04Nb-

0.1C+0.5Mo/Cr steels during deformation and ageing in the ferrite. The chemical

composition of the three steel is listed in Chapter 3.2 and the aTMP applied is

described in Chapter 3.3.4. The microstructural characterisation methods can be

found in Chapter 3.5 and the hardness testing is described in Chapter 3.6.1.

7.2 Results

7.2.1 Hardness

The Vickers hardness of all three steels as a function of ageing time at 600°C is

shown in Figure 7.1. As expected, in all cases, the initial value of the hardness is

high due to the work imparted during the warm rolling. The hardness then

decreases, and this is expected to be a competition between the recovery


104

processes that soften the material and the further solute clustering and

precipitation processes that occur during ageing [30], which harden the material.

For example, the base Nb-containing steel exhibits a hardness in the as-deformed

state of 253±2 HV5, which decreases to 227±2 HV5 after ageing at 10 min ageing.

The Nb-Cr steel is slightly harder at 260±2 HV5 in the as-deformed condition. This

steel also undergoes softening during ageing at 600 °C, reaching 232±4 HV5 after

60 min. The Nb-Mo steel was the hardest, having a hardness of 319±4 HV5 in

as-deformed condition and diminishing relatively less, to 292±3 HV5 after 60 min

ageing at 600°C. In this case, the hardness remains approximately the same at

ageing for 10 min and 30 min. For high-resolution characterisation of the Nb-Mo

steel the sample with 30 min ageing time was chosen. The ageing time of 10 min

was selected for high-resolution characterisation of the Nb-Cr steel since this was

the comparable hardest state, during post-deformation ageing.

Figure 7.1 - Evolution of hardness during ageing at 600°C for the three experimental HSLA

steels warm-deformed in the ferrite.


105

7.2.2 Dislocations and interfaces

The SEM imaging presented here was done in the centre region of the deformed

samples as no significant difference between surface and centre was observed.

Generally, it was found that in all conditions the grains are elongated in rolling

direction (RD), as expected. Figure 7.2 shows a qualitative schematic of the

microstructural evolution of all three alloys as well as some representative

examples of the microstructures in ECCI contrast (ND indicates the normal

direction). The fraction of recovery/recrystallisation was qualitatively obtained from

the ECCI images. Recrystallised grains are individual cells with uniform contrast

within the cell whereas deformed areas show a smeared contrast in the ECCI

images. In the as-deformed conditions (0 min ageing time), the main fraction of

the microstructure in all three steels is dominated by highly deformed areas, as

can be seen in the ECCI images as blurry contrast and as highlighted red areas

in Figure 7.2a-c. Here, only a small amount of recovery and subgrain formation

was generally observed. Upon ageing, distinct changes in grain shape and

morphology can be seen. After 10 min ageing, the Nb-Cr and Nb-Mo steels still

exhibit a large fraction of as-deformed areas whereas in the Nb steel grains appear

to be more equiaxed. Here, subgrains and some new grains have formed (marked

as blue areas in Figure 7.2a-c). In the Nb-Cr steels, the formation of distinct

subgrains can be seen after 10 min ageing time, however, the majority of the

microstructure remains in the as-deformed state. With increasing ageing time up

to 60 min, the ratio of recovered areas with distinct subgrains to as-deformed areas

increases, and the formation of more globular new grains can be observed. This

is in contrast to the Nb-Mo steel where grains remain elongated and as-deformed

even after ageing for 60 min. Only a small amount of subgrains, especially near

the original grain boundaries is observed in this steel. The occasional bright spots

are cementite particles which will be studied further in the following.

Figure 7.3 shows examples of typical EBSD inverse pole figure (IPF) maps for all

three steel in as-deformed and 10 min ageing condition, with high angle grain

boundaries (>15°) shown as black lines and low angle grain boundaries (2-15°)

as white lines. Generally, all three steel show a typical body-centred cubic rolling

texture with grains oriented in <001> and <111> parallel to ND. However, a

specific difference in the recovery mechanism can be seen. The Nb and Nb-Cr

steels only exhibit a small amount of small new grains that form adjacent to the


106

original grain boundaries (marked with white arrows in Figure 7.3). This is in

contrast to the Nb-Mo steel where a large amount of these new grains is seen (see

zoomed-in region in Figure 7.3f). After 10 min ageing these small grains undergo

coarsening in the Nb and Nb-Cr steel but remain stable in the Nb-Mo steel. The

average grain size measured from EBSD for the Nb steel is 1.37±0.05 µm in as-

deformed condition and 1.44±0.08 µm after 10 min ageing. The grain sizes for the

Nb-Cr and Nb-Mo steel in as-deformed condition are 0.95±0.12 µm and

0.56±0.02 µm, after 10 min 0.97±0.05 µm and 0.49±0.06 µm and after 60 min

ageing (not shown here) 1.16±0.06 µm and 0.47±0.04 µm, respectively.

Crystallites confined either by high angle or low angle grain boundaries were

considered for the grain size measurement. Overall, the ratio of high angle grain

boundaries to low angle grain boundaries was roughly 50:50 for all three steels in

their as-deformed condition, and changed slightly towards ~40:60 with an increase

in ageing time due to additional static recovery processes. No significant

difference was found between the individual steels.


107

Figure 7.2 - a) ECCI example images Nb as-deformed, b) ECCI example image Nb-Mo

30 min, c) ECCI example image Nb-Cr 60 min; red areas indicating examples of

as-deformed areas and blue areas indicating examples subgrains, d) Qualitative evolution

of the microstructure of all three alloys over time.


108

Figure 7.3 - Example EBSD IPF maps of all three steels in as-deformed condition a) Nb,

b) Nb-Cr and c) Nb-Mo and in 10 min ageing condition d) Nb, e) Nb-Cr and f) Nb-Mo with

inset highlighting the formation of new grains and subgrains. High angle grain boundaries

(>15°) are shown as black lines and low angle grain boundaries (2-15°) are white lines.

The white arrows highlight exemplary small recrystallised grains near original grain

boundaries. The inset shows IPF colouring in normal direction (vertical).


109

7.2.3 Solute clusters and second phase precipitates

High-resolution characterisation of the solute atom clusters and precipitates was

performed using TEM and AP to study their location within the matrix, their

chemical composition as well as their morphology. The samples studied were

those that exhibited the highest hardness during the post-deformation ageing step.

This was the 10 min ageing condition for the Nb-Cr sample, and the 30 min ageing

condition for the Nb-Mo sample, Figure 7.1. In Figure 7.4, BF-STEM images and

energy-dispersive X-ray spectroscopy (EDS) mapping of the Nb-Cr sample are

shown. It can be seen that there are two types of precipitates: elongated 'larger'

precipitates (100 to 300 nm) that nucleate preferably at the grain boundaries, and

'smaller' (10 to 30 nm), spheroidal intragranular precipitates. EDS mapping

indicated that the larger precipitates were rich in Fe-Mn-Cr-C, whereas the smaller

precipitates contained only Nb and C.

Figure 7.5 shows the STEM and EDS mapping from the Nb-Mo sample after

30 min ageing. Similar to the observations from the Nb-Cr steel, two types of

precipitates were found, 'larger' ones at grain boundaries and 'smaller'

intragranular precipitates. In accordance with the ECCI, there was a higher amount

of dislocation debris present compared to the Nb-Cr sample, as seen in Figure

7.5c. The EDS mapping reveals that the 'larger' precipitates are Fe-Mn-C-Mo rich

whereas the 'smaller' precipitates are rich in Nb and Mo.


110

Figure 7.4 - BF-STEM images, ADF-STEM and EDS mappings of Nb-Cr steel after 10

min ageing. a) Precipitates at the grain boundary, b) precipitates inside the grains, c) EDS

elemental mappings of grain boundary precipitates (also showing some small in-grain

precipitates), d) ADF-STEM and EDS elemental mappings of in-grain precipitates only.

The arrows highlight exemplary precipitates at the grain boundary (Figure 7.4a) and inside

the grain (Figure 7.4b), respectively.


111

Figure 7.5 - BF-STEM images, ADF images and EDS mappings of Nb-Mo steel after

30 min ageing, a) larger precipitates at the grain boundary (as indicated by arrows), b)

smaller, spheroidal intragranular precipitates (arrows), c) dislocation networks (arrows),

d) ADF image and EDS elemental mappings of grain boundary precipitates, e) ADF image

and EDS elemental mappings of in-grain precipitates.

For precipitate size evaluation, at least 100 precipitates per type and sample were

measured from TEM images. It can be said that most precipitates at grain

boundaries are FeMnC-rich and that most precipitates inside the grains are NbC.

Figure 7.6a shows the precipitate size (diameter) of particles at grain boundaries.

The Nb-Cr and Nb-Mo samples have similar values of around 143 nm. However,

please note that the Nb-Mo sample was aged for 30 min. The average precipitate

size of the Nb steel aged for 10 min is around 211 nm which is considerably larger

as compared to the Nb-Cr steel at the same ageing time. In Figure 7.6b, the

precipitate sizes inside the grains can be seen which is very similar with around

25 nm for all three steels.


112

Figure 7.6 - Precipitate size distributions for the experimental steels aged at 600 °C: the

Nb-Cr steel aged for 10 min, and the Nb-Mo steel aged for 30 min. The sizes were

measured from TEM images. a) 'Larger' precipitates at the grain boundaries, and b)

'Smaller' intragranular precipitates.

To further study the chemical composition of the precipitates and to search for

nanoscale clusters, AP was performed on the same samples that were studied

with TEM.

Figure 7.7 provides atom maps for C, Nb, Mo and Nb-C and includes a grain

boundary, as well as a precipitate. The C segregation along a vertical line in the C

reconstruction which is marked with a blue arrow in Figure 7.7 is an artefact from

preferential C evaporation at a crystallographic pole, which is an area with a high

field gradient causing the diffusion of the C atoms during AP measurements [83].

A region of interest was inserted through a selected precipitate, as indicated by

the cylinder 1 in Figure 7.7a. This was used to obtain the 1D concentration profile

shown in Figure 7.7b, with the arrow marking start and finish position of this profile.

Figure 7.7b shows a depletion of Fe and an increase in elements such as Nb, Mo

and C whereas Mo was only present with a concentration of around 10 at.%.

Another region of interest was inserted through an interface, as indicated by

cylinder 2 in Figure 7.7a, and this reveals that Mn, C, Mo and Nb segregates to

this interface. The arrow marks the start and finish position of the corresponding

1D concentration profile shown in Figure 7.7c. The cylinder did not enter in the

pole region. This profile shows that Mn and C, followed by Mo have a higher

tendency for grain boundary segregation as compared to Nb.


113

Figure 7.7 - AP reconstruction of Nb-Mo 30 min (total number of atoms: 39.6 million)

showing a) elemental distribution of C, Nb, Mo and the region of interest through the

precipitate and grain boundary, b) 1D concentration profile precipitate based on region of

interest inserted as cylinder 1, c) 1D concentration profile through a boundary based on

region of interest as cylinder 2; the blue arrow indicates a pole; the black arrows indicate

start and finish of the concentration profiles.


114

The AP tomograms in Figure 7.8 reveal the atomic co-clustering of Nb and C in

the Nb-Mo steel aged 30 min and the Nb-Cr steel aged for 10 min. These atomic

cluster maps were generated using a maximum separation distance of dmax= 1.33

nm, and a minimum number of atoms in a cluster of Nmin= 6 for the datasets in

Figure 7.8a. Similarly, values of dmax= 1.01 and Nmin= 10 for the dataset shown in

Figure 7.8b were used. Figure 7.8c and d summarise the number densities for the

Nb-Mo steel and the Nb-Cr steel, respectively. In the following, clusters will be

defined as features containing <60 atoms and the term nano-precipitates used to

describe features containing >60 atoms.

For the Nb-Mo steel, the majority of the Nb-C co-clusters observed contained

<20 atoms. However, a second peak of nano-precipitates was observed with

>60 atoms per cluster. Furthermore, it was found that these nano-precipitates

incorporate a significant amount of Mo after reaching a diameter of ~5 nm,

equivalent to them containing ≥ 4 atoms.

Figure 7.8d shows Nb-C co-clusters from the Nb-Cr containing steel. Relatively, a

much lower number of nano-precipitates were observed. Moreover, Cr was not

incorporated into the clusters, nor into the nano-precipitates.

The local matrix composition of the reconstruction shown in Figure 7.8a is

Fe-0.01C-1.01Mn-0.02Nb-0.55Mo wt.% and for the reconstruction in Figure 7.8b

is Fe-0.02C-1.1Mn-0.03Nb-0.64Cr wt.%. This suggests that the matrix is depleted

in Mn, C and Nb as compared to the overall steel compositions shown in Table

3.2.


115

Figure 7.8 - AP reconstruction showing NbC clusters a) Nb-Mo 30 min and b) Nb-Cr

10 min cluster particle density over cluster size for c) Nb-Mo 30 min, d) Nb-Cr 10 min.

Nb

C


116

7.3 Discussion

In this chapter, the influence of Mo and Cr as microalloying elements in a

Nb-based HSLA steel which was deformed and aged in the ferrite was

investigated. In the following section, the effect of Cr/Mo on dislocations, solid

solution, clusters and precipitates as well as interfaces will be discussed

separately. A comprehensive model of the microstructural mechanisms will be

introduced at the end.

7.3.1 Dislocations and interfaces

The as-deformed microstructures of all three alloys consist of as-deformed areas

and recovered areas with subgrain structures as well as new recrystallised grains

(Figure 7.2 and Figure 7.3). Due to the high stacking fault energy of ferrite, the

main mechanism behind this is continuous dynamic recrystallisation, where there

is an extended tendency for dislocation rearrangements/annihilation and subgrain

formation. Subgrain boundaries can convert into high angle grain boundaries if

sufficient strain is applied [68]. The addition of Cr, as well as Mo, obviously leads

to a retardation in continuous dynamic recrystallisation as compared to the base

alloy which only contains Nb (Figure 7.2 and Figure 7.3). However, Mo is more

effective in retarding dislocation annihilation and subgrain formation as compared

to Cr and Nb as can be seen in Figure 7.2, Figure 7.3 and Figure 7.5c. This

microstructural observation corresponds well with the hardness data shown in

Figure 7.1 where the hardness of the Nb-Mo steel is ~20% higher as compared to

the Nb-Cr and Nb steels. It is suggested that the solute drag effect of Mo [18] is a

major contributor in delaying the dislocation rearrangements/annihilation and

subgrain formation. The high amount of dislocation networks present in the Nb-

Mo alloy will, thus, act as nucleation sites for NbC and also contribute to

strengthening. This is also suggested by Uemori et al. [25] who studied the

addition of Mo in a Ti-Nb steel via AP field ion microscopy as well as in the review

on Mo alloying in HSLA steels by Mohrbacher [18]. Furthermore, the segregation

of Mo as well as Nb and Mn to the grain boundary, as seen in Figure 7.7, may be

able to pin the grain boundary and, thus, may provide hardening. Mo segregation

in conjunction with P segregation to the prior austenite boundaries has been

reported by Song et al. [116] for a Cr-Mo low-alloyed steel that has been aged at

540°C after quenching from 980°C.


117

7.3.2 Precipitates and clusters

Mo is incorporated into both NbC and FeMnC as can be seen from Figure 7.5,

Figure 7.7 and Figure 7.8. The chemical composition of the matrix for the AP

reconstruction shown in Figure 7.8a clearly suggests that the majority of Mo

remains in solid solution, which contributes to solid solution strengthening. Only

small amounts of Mo participate in the NbC formation rather than segregating at

the precipitate/matrix interface as suggested in the literature by Cao et al. [26] and

Enloe et al. [27]. Cao et al. [26] studied a 0.08Nb-0.14Mo microalloyed steel after

warm deformation at 880°C, where complex (Nb,Mo)(C,N) with a ratio of Nb:Mo

of 2.43 were formed. Similar observations have been made for TiMo-

steels [27,29,30,117]. Enloe et al. [27] found Mo incorporated in NbC at 900°C

using AP. They also suggest that Mo is expected to be more significantly

incorporated into NbC in the ferrite compared to NbC in austenite based on an

ideal solution model after Speer et al. [118]. Nöhrer et al. [119] studied the

evolution of Nb precipitates in a low-alloyed steel which was deformed at 700°C.

They reported that Nb precipitates have a higher volume fraction in ferrite

compared to the same processing in austenite due to a lower solubility and higher

diffusivity in ferrite. On the contrary, Park et al. [31] did not find Mo addition in NbC

precipitates of a 0.45 wt.% Nb and 0.19 wt.% Mo microalloyed steel which was hot

rolled and then aged at temperatures between 500-700°C but found the Mo rather

in the matrix.

To understand why Mo is incorporated in carbides, Jang et al. [120] calculated the

lattice parameters of MoC with first principle showing that the replacement of Ti by

Mo will reduce the misfit with ferrite, which facilitates nucleation. At the same time,

the substitution of Ti by Mo will reduce the equilibrium Ti concentration in the ferrite

matrix during coarsening, which possibly decelerates the coarsening process of

(Ti,M)C precipitation. Mo retards growth of precipitates due to its partition into the

matrix as it is not favoured within the TiC. This behaviour is similar for Mo in

NbC/γ-Fe as calculated with density functional theory by Zhou et al. [121]. Nb

replacement by Mo is energetically unfavourable, however, it has been reported

to reduce the lattice parameter of MC and interfacial chemical energy which makes

it more resistant to coarsening compared to pure NbC [121]. To our best

knowledge, no first principle studies on Cr incorporation into NbC have been

reported in literature.


118

Overall, it seems that Mo is more effective in refining the precipitate size as

compared to Cr, (Figure 7.6) which also suggests that a combination of higher

number of nucleation sites and retarded diffusivity of Nb and C may be responsible

for finer size and increased hardness. Kostryzhev et al. [20] reported that Mo has

a stronger influence than Cr on the retardation of cementite coarsening in a NbV

microalloyed steel.

The cluster analysis (Figure 7.8) shows that in both alloys NbC clusters are

present. Mo is only present in clusters/nano-precipitates with a diameter larger

than ~5 nm. A similar behaviour was observed by others in TiMo steels [30,122]

suggesting that Mo is not only a key element in controlling the precipitate growth

but also in the early stages of precipitation.

Figure 7.4 shows that Cr is incorporated in the larger FeMn-rich precipitates in the

Nb-Cr steel in contrary to the nanoscale NbC where no Cr was found. A similar

behaviour of Cr was also observed by Kostryshev et al. [20] for a NbV-

microalloyed bainitic steel where it was found that Cr was effective in retarding the

growth of FeMnC-rich precipitates. This too can be observed in Figure 7.6a in the

present study. The average size of the FeMnC-rich precipitates in this study after

10 min ageing for the Nb-Cr steel is around a 30% finer compared to the Nb steel.

For the Nb-Cr sample no significant contribution of Cr in the cluster/nano-

precipitates was observed. However, Pereloma et al. [123] reported Cr clusters in

a low C steel, however, in this steel no Nb was present. Xie et al. [124] reported

NbC clusters in a 0.084 wt.% Nb microalloyed steel that did contain 0.37 wt.% Cr.

In their study, there was no evidence that NbC was enriched in Cr.

7.3.3 Microstructural Model

A schematic of the microstructural features upon Cr/Mo addition is shown in Figure

7.9 to summarise the detailed effects of Mo versus Cr. Both steels, the Nb-Cr as

well as the Nb-Mo, exhibit grains that are subdivided by low angle grain boundaries

and new recrystallised grains. FeMnC-rich precipitates are found at original grain

boundaries whereas NbC-rich precipitates and clusters are preferably located

inside of grains. The main difference between these two steels is that Cr only

partitions into the FeMnC-rich precipitates whereas Mo partitions into the FeMnC-

rich precipitates as well as into NbC. Furthermore, Mo retards dislocations

annihilation and subgrain formation, hence, small recrystallised grains remain


119

stable upon ageing. NbC clusters were found in the Nb-Cr as well as in the Nb-Mo

steels. Interestingly, upon reaching a diameter of roughly 5 nm Mo additions in the

NbC clusters in the Nb-Mo steel were found.

Figure 7.9 - Schematic of the microstructural features observed upon Cr/Mo addition in

HSLA steels subjected to warm deformation in the ferrite.

Cr as microalloying element is, thus, not effective in contributing to strengthening

during TMP processing of HSLA steels in the ferrite. However, it is assumed that

Cr still provides corrosion resistance although this was not investigated here. Mo,

on the other hand, leads to a significant increase in strengthening due to the

retardation in grain growth and precipitate coarsening. Therefore, it is suggested

that a combination of Nb and Mo as microalloying elements in HSLA steels

subjected to deformation in the ferrite is a favourable combination to achieve

increased yield strengths. In comparison to previous studies of some of the current

authors [90,107] where a Nb-Mo-Ti microalloyed steel was subjected to warm

deformation, it can be concluded that an alloying concept without Ti is more

advantageous as Nb remains in solid solution and does not diffuse to pre-existing

TiN. Therefore Nb is available for strain-induced precipitation.


120

7.4 Summary

In this chapter, the influence of Cr and Mo as microalloying elements on the

microstructural evolution of a Fe-1.6Mn-0.04C-0.1Nb model alloy subjected to

deformation in the ferrite region was investigated. The main findings are

summarised in the following:

Mo provides a significant increase in hardness of ~20% compared to the

base material (without Mo) whereas the Cr addition provides only a minor

hardening effect.

It is shown that Mo is more effective in contributing to strengthening due to

its partitioning into FeMnC-rich precipitates as well as into NbC precipitates

and clusters so as to inhibit precipitate growth during ageing.

Mo, Nb and Mn were found to segregate to interfaces/grain boundaries in

the Nb-Mo steel and contribute to strengthening by reducing the mobility of

these interfaces.

Mo retards recovery and the processes of dislocation annihilation and

subgrain formation due to a drag effect during ageing at 600°C.

Cr, on the other hand, partitions into the FeMnC-rich precipitates

exclusively and slows down their coarsening, but not does not provide the

additional beneficial effects of Mo summarised above. It was not found in

small NbC precipitates and clusters.

Conclusion and Outlook ________________________________________________________________________________________________________________________________

121

8 Conclusion and Outlook

Conclusions of the project objectives of advanced thermo-mechanical processing

of modern HSLA steels outlined in Chapter 1 are reviewed below. Overall, it can

be concluded that hierarchical microstructures can be achieved in HSLA steels

subjected to warm deformation in the ferrite. However, more work needs to be

done to optimise the process and alloy design. The overall findings have been

summarised in Table 8.1: Summary of findings Recommendations for future research

aims are discussed at the end.

Table 8.1: Summary of findings

microstructure degree of homogeneity hardness

Chapter 4 -

feasibility

good bad good

Chapter 5 -

optimisation

good good medium

Chapter 6 –

scale up

medium good medium

Chapter 7 –

alloy design

good good good

8.1 HSLA steels and aTMP

Hierarchical microstructure can be achieved for an industrial Ti-Nb-Mo

microalloyed HSLA steel with a martensitic/bainitic starting microstructure

subjected to warm deformation at 600°C in a Gleeble. These microstructures

consist of ultrafine crystallites that are confined by a combination of high angle and

low angle grain boundaries that are decorated by two types of precipitates. Large

FeMnC-rich cementite particles precipitate on grain boundaries and smaller

TiNbC-rich precipitates nucleate on dislocations and subgrain boundaries. Single

pass deformation at a strain rate of 10 s-1 leads to a macroscopic shear band

formation, therefore, it was essential to develop a aTMP route that is optimised for

the steel investigated in this study. A soaking time of 10 min at 600°C prior to


122

warm deformation, as well as a three-pass deformation with a strain of 0.2 per

pass with a strain rate of 1 s-1 allows to achieve a homogenous deformation during

plane strain compression testing.

8.2 Mechanical properties

Rolling at a temperature around 650°C with a reduction of thickness of 55% results

in a UTS of 650 MPa, a yield ratio of 0.95 and a total elongation of 14%. The low

yield ratio can be explained by the rather large precipitates (average precipitate

size for the as-rolled condition: 51±38 nm) in the Ti-Nb-Mo microalloyed steel,

which do not effectively contribute to work hardening. In the lower temperature

region of the Charpy impact testing delaminations occurred in longitudinal as well

as in transversal directions. However, direct ageing for 60 min shows a very high

impact energy at room temperature due to the onset of grain growth during

recrystallisation.

8.3 Alloy design

Additions of 0.5wt% Mo lead to an increase in hardness of ~20% compared to a

plain Nb microalloyed steel, whereas 0.5wt% Cr provides only a minor addition to

hardening. It was found that Mo is more effective than Cr in delaying dislocation

recovery and that it partitions into nanoscale Nb-C solute clusters and precipitates

of NbC and Fe3C during ageing, retarding the coarsening of these phases.

Conversely, Cr was found to solely partition into Fe3C, and does not contribute to

the nature of the dispersion of clusters and NbC.

8.4 Recommendations for future research

As TiN is already formed during casting and is not dissolved at typical

austenitisation temperatures Nb tends to co-precipitate at these TiN and

hence is not available for strain induced precipitation. Therefore, it is

suggested to focus on an alloy design without Ti addition to obtain a smaller

precipitate size and a higher work hardening rate.

An increase in applied strain might be beneficial to obtain more

recrystallised grains and hence a lower DBTT.


123

The influence of heating rate could be explored in more detail. In

Chapter 2.5.2 it was discussed that a larger reheating rate can lead to finer

precipitates. Therefore, it might be beneficial to carefully select the heating

rate applied depending on the chemical composition of the steel.

Further, it is suggested to more carefully study the impact toughness in the

area of the DBTT and compare the results with a coarse grained steel.

As a combination of Nb and Mo as microalloying elements has been shown

to be promising in achieving exceptional strengths steels subjected to warm

deformation, it is suggested to upscale and study the mechanical

properties.

It might be interesting to explore further additions of microalloying elements

such as B or Cu. B in conjunction with Mo has been reported to form Mo-B

cluster and hence leading to a decrease in C and N diffusion as discussed

in Chapter 2.3. Cu has been reported to increase corrosion resistance (see

Chapter 2.2) and also forms Cu clusters. These clusters have been

reported to contribute to hardening in precipitation hardening

steels [125,126].

References ________________________________________________________________________________________________________________________________

124

9 References

[1] Worldsteel Association, World Steel in Figures 2019, (2019). https://www.worldsteel.org/en/dam/jcr:96d7a585-e6b2-4d63-b943-4cd9ab621a91/World%2520Steel%2520in%2520Figures%25202019.pdf (accessed August 3, 2019).

[2] Worldsteel Association, Steel Markets, (2019). https://www.worldsteel.org/steel-by-topic/steel-markets.html (accessed August 3, 2019).

[3] Worldsteel Association, Steel in Buildings and Infrastructure, (2019). https://www.worldsteel.org/steel-by-topic/steel-markets/buildings-and-infrastructure.html (accessed August 14, 2019).

[4] R. Song, D. Ponge, D. Raabe, J.G. Speer, D.K. Matlock, Overview of processing, microstructure and mechanical properties of ultrafine grained bcc steels, Mater. Sci. Eng. A. 441 (2006) 1–17. doi:10.1016/j.msea.2006.08.095.

[5] A. Ohmori, S. Torizuka, K. Nagai, Strain-hardening due to Dispersed Cementite for Low Carbon Ultrafine-grained Steels, ISIJ Int. 44 (2004) 1063–1071. doi:10.2355/isijinternational.44.1063.

[6] Y. Okitsu, N. Takata, N. Tsuji, A new route to fabricate ultrafine-grained structures in carbon steels without severe plastic deformation, Scr. Mater. 60 (2009) 76–79. doi:10.1016/j.scriptamat.2008.09.002.

[7] B. Eghbali, Effect of strain rate on the microstructural development through continuous dynamic recrystallization in a microalloyed steel, Mater. Sci. Eng. A. 527 (2010) 3402–3406. doi:10.1016/j.msea.2010.01.077.

[8] L. Cheng, Y. Chen, Q. Cai, W. Yu, G. Han, E. Dong, X. Li, Precipitation enhanced ultragrain refinement of Ti-Mo microalloyed ferritic steel during warm rolling, Mater. Sci. Eng. A. 698 (2017) 117–125. doi:10.1016/j.msea.2017.04.112.

[9] R. Song, D. Ponge, D. Raabe, Mechanical properties of an ultrafine grained C-Mn steel processed by warm deformation and annealing, Acta Mater. 53 (2005) 4881–4892. doi:10.1016/j.actamat.2005.07.009.

[10] M. Koyama, Z. Zhang, M. Wang, D. Ponge, D. Raabe, K. Tsuzaki, H. Noguchi, C.C. Tasan, Bone-like crack resistance in hierarchical metastable nanolaminate steels, Science (80-. ). 355 (2017) 1055–1057. doi:10.1126/science.aal2766.

[11] Y.M. Wang, T. Voisin, J.T. McKeown, J. Ye, N.P. Calta, Z. Li, Z. Zeng, Y. Zhang, W. Chen, T.T. Roehling, R.T. Ott, M.K. Santala, P.J. Depond, M.J. Matthews, A. V. Hamza, T. Zhu, Additively manufactured hierarchical stainless steels with high strength and ductility, Nat. Mater. 17 (2018) 63–70. doi:10.1038/NMAT5021.

[12] R. Song, D. Ponge, D. Raabe, Improvement of the work hardening rate of ultrafine grained steels through second phase particles, Scr. Mater. 52 (2005) 1075–1080. doi:10.1016/j.scriptamat.2005.02.016.

[13] A.J. DeArdo, M. Hua, K. Cho, C.I. Garcia, On the strength of microalloyed steels - An interpretive review, Mater. Sci. Technol. 25 (2009) 1074–1082.

References ________________________________________________________________________________________________________________________________

125

doi:10.1179/174328409X455233. [14] T.N. Baker, Microalloyed Steels, Ironmak. Steelmak. 43 (2016) 264–307.

doi:10.1179/1743281215Y.0000000063. [15] J. Takahashi, K. Kawakami, J.I. Hamada, K. Kimura, Direct observation of

niobium segregation to dislocations in steel, Acta Mater. 107 (2016) 415–422. doi:10.1016/j.actamat.2016.01.070.

[16] S. Vervynckt, K. Verbeken, B. Lopez, J.J. Jonas, Modern HSLA steels and role of non-recrystallisation temperature, Int. Mater. Rev. 57 (2012) 187–207. doi:10.1179/1743280411Y.0000000013.

[17] M.G. Akben, B. Bacroix, J.J. Jonas, Effect of Vanadium and Molybdenum Addition on High Temperature Recovery, Recrystallization and Precipitation Behavior of Niobium-Based Microalloyed Steels, 31 (1983) 161–174.

[18] H. Mohrbacher, Principal effects of Mo in HSLA steels and cross effects with microalloying elements, in: Int. Semin. Appl. Mo Steels, Beijing, China, 2010.

[19] T. Jia, M. Li, X. Pei, Z. Wang, On the Spheroidizing Annealing Behavior in Cr/Nb Microalloyed Medium Carbon Steels, Steel Res. Int. 1800353 (2018) 1–11. doi:10.1002/srin.201800353.

[20] A. Kostryzhev, N. Singh, L. Chen, C. Killmore, E. Pereloma, Comparative Effect of Mo and Cr on Microstructure and Mechanical Properties in NbV-Microalloyed Bainitic Steels, Metals (Basel). 8 (2018) 134. doi:10.3390/met8020134.

[21] W.B. Morrison, Microalloy steels - the beginning, Mater. Sci. Technol. 25 (2009) 1066–1073. doi:10.1179/174328409x453299.

[22] J. Takahashi, K. Kawakami, J.I. Hamada, K. Kimura, Direct observation of niobium segregation to dislocations in steel, Acta Mater. 107 (2016) 415–422. doi:10.1016/j.actamat.2016.01.070.

[23] M.G. Akben, B. Bacroix, J.J. Jonas, Effect of Vanadium and Molybdenum Addition on High Temperature Recovery, Recrystallization and Precipitation Behavior of Niobium-Based Microalloyed Steels, Acta Metall. 31 (1983) 161–174. https://ac.els-cdn.com/0001616083900767/1-s2.0-0001616083900767-main.pdf?_tid=44db01b4-9d14-4b4c-ad2c-677592dbe09b&acdnat=1544116170_fd17d67963e0acf8c1a766cc01d051cd.

[24] W.B. Lee, S.G. Hong, C.G. Park, S.H. Park, Carbide precipitation and high-temperature strength of hot-rolled high-strength, low-alloy steels containing Nb and Mo, Metall. Mater. Trans. A Phys. Metall. Mater. Sci. 33 (2002) 1689–1698. doi:10.1007/s11661-002-0178-2.

[25] R. Uemori, R. Chijiiwa, H. Tamehiro, H. Morikawa, AP-FIM study on the effect of Mo addition on microstructure in Ti-Nb steel, Appl. Surf. Sci. 76–77 (1994) 255–260. doi:10.1016/0169-4332(94)90351-4.

[26] J. Cao, Q. Yong, Q. Liu, X. Sun, Precipitation of MC phase and precipitation strengthening in hot rolled Nb-Mo and Nb-Ti steels, J. Mater. Sci. 42 (2007) 10080–10084. doi:10.1007/s10853-007-2000-4.

[27] C.M. Enloe, K.O. Findley, C.M. Parish, M.K. Miller, B.C. De Cooman, J.G. Speer, Compositional evolution of microalloy carbonitrides in a Mo-bearing microalloyed steel, Scr. Mater. 68 (2013) 55–58. doi:10.1016/j.scriptamat.2012.09.027.

[28] J. Lu, O. Omotoso, J.B. Wiskel, D.G. Ivey, H. Henein, Strengthening

References ________________________________________________________________________________________________________________________________

126

mechanisms and their relative contributions to the yield strength of microalloyed steels, Metall. Mater. Trans. A Phys. Metall. Mater. Sci. 43 (2012) 3043–3061. doi:10.1007/s11661-012-1135-3.

[29] J.H. Jang, Y.-U. Heo, C.-H. Lee, H.K.D.H. Bhadeshia, D.-W. Suh, Interphase precipitation in Ti–Nb and Ti–Nb–Mo bearing steel, Mater. Sci. Technol. 29 (2013) 309–313. doi:10.1179/1743284712Y.0000000131.

[30] S. Mukherjee, I. Timokhina, C. Zhu, S.P. Ringer, P.D. Hodgson, Clustering and precipitation processes in a ferritic titanium-molybdenum microalloyed steel, J. Alloys Compd. 690 (2017) 621–632. doi:10.1016/j.jallcom.2016.08.146.

[31] D.B. Park, M.Y. Huh, J.H. Shim, J.Y. Suh, K.H. Lee, W.S. Jung, Strengthening mechanism of hot rolled Ti and Nb microalloyed HSLA steels containing Mo and W with various coiling temperature, Mater. Sci. Eng. A. 560 (2013) 528–534. doi:10.1016/j.msea.2012.09.098.

[32] T. Hara, H. Asahi, R. Uemori, H. Tamehiro, Role of Combined Addition of Niobium and Boron and of Molybdenum and Boron on Hardnenability in Low Carbon Steels, ISIJ Int. 44 (2008) 1431–1440. doi:10.2355/isijinternational.44.1431.

[33] M.A. Razzak, Heat treatment and effects of Cr and Ni in low alloy steel, Bull. Mater. Sci. 34 (2011) 1439–1445. doi:10.1007/s12034-011-0340-9.

[34] E. V Pereloma, V. Bata, R.I. Scott, R.M. Smith, Effect of Cr on Strain Ageing Behaviour of Low Carbon Steel, (2007) 4214–4219. doi:10.4028/www.scientific.net/MSF.539-543.4214.

[35] I.B. Timokhina, P.D. Hodgson, S.P. Ringer, R.K. Zheng, E. V. Pereloma, Precipitate characterisation of an advanced high-strength low-alloy (HSLA) steel using atom probe tomography, Scr. Mater. 56 (2007) 601–604. doi:10.1016/j.scriptamat.2006.12.018.

[36] S.P. Tsai, T.C. Su, J.R. Yang, C.Y. Chen, Y.T. Wang, C.Y. Huang, Effect of Cr and Al additions on the development of interphase-precipitated carbides strengthened dual-phase Ti-bearing steels, Mater. Des. 119 (2017) 319–325. doi:10.1016/j.matdes.2017.01.071.

[37] E.O. Hall, The Deformation and Ageing of Mild Steel: III Discussion of Results, Proc. Phys. Soc. Sect. B. 64 (1951) 747–753. doi:10.1088/0370-1301/64/9/303.

[38] L.. Teoh, Thermo-mechanical processing and microstructure of microalloyed steel bar and wire rod products, J. Mater. Process. Technol. 48 (1995) 475–481.

[39] N. Isasti, D. Jorge-Badiola, M.L. Taheri, P. Uranga, Microstructural and precipitation characterization in Nb-Mo microalloyed steels: Estimation of the contributions to the strength, Met. Mater. Int. 20 (2014) 807–817. doi:10.1007/s12540-014-5002-1.

[40] Y.W. Kim, S.W. Song, S.J. Seo, S.G. Hong, C.S. Lee, Development of Ti and Mo micro-alloyed hot-rolled high strength sheet steel by controlling thermomechanical controlled processing schedule, Mater. Sci. Eng. A. 565 (2013) 430–438. doi:10.1016/j.msea.2012.12.055.

[41] Y.W. Kim, J.H. Kim, S.G. Hong, C.S. Lee, Effects of rolling temperature on the microstructure and mechanical properties of Ti-Mo microalloyed hot-rolled high strength steel, Mater. Sci. Eng. A. 605 (2014) 244–252. doi:10.1016/j.msea.2014.03.054.

[42] S. Ghosh, S. Mula, Thermomechanical processing of low carbon Nb-Ti

References ________________________________________________________________________________________________________________________________

127

stabilized microalloyed steel: Microstructure and mechanical properties, Mater. Sci. Eng. A. 646 (2015) 218–233. doi:10.1016/j.msea.2015.08.072.

[43] B.Q. Han, S. Yue, Processing of ultrafine ferrite steels, J. Mater. Process. Technol. 136 (2003) 100–104. doi:10.1016/S0924-0136(02)01014-2.

[44] E. Astafurova, G. Maier, E. Melnikov, E. Naydenkin, A. Smirnov, V. Bataev, P. Odessky, S. Dobatkin, H.J. Maier, The Influence of the Thermomechanical Processing Regime on the Structural Evolution of Mo-Nb-Ti-V Microalloyed Steel Subjected to High-Pressure Torsion, Metall. Mater. Trans. A Phys. Metall. Mater. Sci. 48 (2017) 3400–3409. doi:10.1007/s11661-017-4085-y.

[45] Y. Kimura, T. Inoue, F. Yin, O. Sitdikov, K. Tsuzaki, Toughening of a 1500 MPa class steel through formation of an ultrafine fibrous grain structure, Scr. Mater. 57 (2007) 465–468. doi:10.1016/j.scriptamat.2007.05.039.

[46] Y. Kimura, T. Inoue, F. Yin, K. Tsuzaki, Inverse Temperature Dependence of Toughness in an Ultrafine Grain-Structure Steel, Science (80-. ). 320 (2008) 1057–1061. doi:10.1126.

[47] U.H. Lee, N. Kamikawa, G. Miyamoto, T. Furuhara, Continuous Dynamic Recrystallization during Warm Deformation of Tempered Lath Martensite in a Medium Carbon Steel, Key Eng. Mater. 508 (2012) 124–127. doi:10.4028/www.scientific.net/KEM.508.124.

[48] R. Song, D. Ponge, D. Raabe, R. Kaspar, Microstructure and crystallographic texture of an ultrafine grained C-Mn steel and their evolution during warm deformation and annealing, Acta Mater. 53 (2005) 845–858. doi:10.1016/j.actamat.2004.10.051.

[49] M. Calcagnotto, Y. Adachi, D. Ponge, D. Raabe, Deformation and fracture mechanisms in fine- and ultrafine-grained ferrite/martensite dual-phase steels and the effect of aging, Acta Mater. 59 (2011) 658–670. doi:10.1016/j.actamat.2010.10.002.

[50] L. Qun, J. Tianfu, G. Yuwei, L. Hongbiao, Investigation on the preparation of the ultrafine grained low-carbon steel by warm deformation, Phase Transitions. 85 (2012) 619–629. doi:10.1080/01411594.2011.651796.

[51] H.F. Lan, W.J. Liu, X.H. Liu, Ultrafine Ferrite Grains Produced by Tempering Cold-rolled Martensite in Low Carbon and Microalloyed Steels, ISIJ Int. 47 (2007) 1652–1657. doi:10.2355/isijinternational.47.1652.

[52] S. Malekjani, I.B. Timokhina, I. Sabirov, P.D. Hodgson, Deformation Behaviour of Ultrafine Grained Steel Produced by Cold Rolling of Martensite, Can. Metall. Q. 48 (2009) 229–235. doi:10.1179/000844309794239017.

[53] Y. Mazaheri, A. Kermanpur, A. Najafizadeh, N. Saeidi, Development of a New Ultrafine/Nano Ferrite-Carbide Microstructure by Thermomechanical Processing, Acta Metall. Sin. (English Lett. 28 (2015) 249–253. doi:10.1007/s40195-014-0191-7.

[54] M. Abbasi, A. Kermanpur, A. Najafizadeh, S. Saeedipour, Y. Mazaheri, The Effect of Niobium on the Formation of Nanostructured Low Carbon Steel Using Martensite Treatment, Int. J. ISSI. 9 (2012) 6–10.

[55] F. Foroozmehr, A. Najafizadeh, A. Shafyei, Effects of carbon content on the formation of nano/ultrafine grained low-carbon steel treated by martensite process, Mater. Sci. Eng. A. 528 (2011) 5754–5758. doi:10.1016/j.msea.2011.04.055.

[56] E. Ghassemali, A. Kermanpur, A. Najafizadeh, Microstructural evolution in

References ________________________________________________________________________________________________________________________________

128

a low carbon steel during cold rolling and subsequent annealing, J. Nanosci. Nanotechnol. 10 (2010) 6177–6181. doi:10.1166/jnn.2010.2578.

[57] H. Azizi-Alizamini, M. Militzer, W.J. Poole, A novel technique for developing bimodal grain size distributions in low carbon steels, Scr. Mater. 57 (2007) 1065–1068. doi:10.1016/j.scriptamat.2007.08.035.

[58] S.M. Hosseini, A. Najafizadeh, A. Kermanpur, Producing the nano/ultrafine grained low carbon steel by martensite process using plane strain compression, J. Mater. Process. Technol. 211 (2011) 230–236. doi:10.1016/j.jmatprotec.2010.09.014.

[59] J. Tian, G. Xu, W. Liang, Q. Yuan, Effect of Annealing on the Microstructure and Mechanical Properties of a Low-Carbon Steel with Ultrafine Grains, Metallogr. Microstruct. Anal. 6 (2017) 233–239. doi:10.1007/s13632-017-0350-0.

[60] Q. Chao, P.D. Hodgson, H. Beladi, Microstructure evolution of martensitic Ti-6Al-4V alloy during warm deformation, Mater. Sci. Forum. 783–786 (2014) 679–684. doi:10.4028/www.scientific.net/MSF.783-786.679.

[61] Q. Chao, P.D. Hodgson, H. Beladi, Ultrafine grain formation in a Ti-6Al-4V alloy by thermomechanical processing of a martensitic microstructure, Metall. Mater. Trans. A Phys. Metall. Mater. Sci. 45 (2014) 2659–2671. doi:10.1007/s11661-014-2205-5.

[62] Z.X. Zhang, S.J. Qu, A.H. Feng, J. Shen, D.L. Chen, Hot deformation behavior of Ti-6Al-4V alloy: Effect of initial microstructure, J. Alloys Compd. 718 (2017) 170–181. doi:10.1016/j.jallcom.2017.05.097.

[63] N. Tsuji, R. Ueji, Y. Minamino, Y. Saito, A new and simple process to obtain nano-structured bulk low-carbon steel with superior mechanical property, Scr. Mater. 46 (2002) 305–310. doi:10.1016/S1359-6462(01)01243-X.

[64] R. Ueji, N. Tsuji, Y. Minamino, Y. Koizumi, Ultragrain refinement of plain low carbon steel by cold-rolling and annealing of martensite, Acta Mater. 50 (2002) 4177–4189. doi:10.1016/S1359-6454(02)00260-4.

[65] B. Poorganji, G. Miyamoto, T. Furuhara, Formation of ultrafine grained ferrite by warm deformation of lath martensite in low-alloy steels with different carbon content, Scr. Mater. 59 (2008) 279–281. doi:10.1016/j.scriptamat.2008.03.041.

[66] T. Furuhara, K. Kobayashi, T. Maki, Control of Cementite Precipitation in Lath Martensite by Rapid Heating and Tempering, ISIJ Int. 44 (2004) 1937–1944. doi:10.2355/isijinternational.44.1937.

[67] J.H. Li, P.G. Xu, Y. Tomota, Y. Adachi, Formation of ferrite-cementite ultrafine grained microstructure by warm compression for SM490 martensite steel, Recryst. Grain Growth III, Pts 1 2. 558–559 (2007) 539–544. doi:10.4028/www.scientific.net/MSF.558-559.539.

[68] S. Gourdet, F. Montheillet, A model of continuous dynamic recrystallization, Acta Mater. 51 (2003) 2685–2699. doi:10.1016/S1359-6454(03)00078-8.

[69] J. Gallego, A.R. Rodrigues, L. Montanari, Second Phase Precipitation in Ultrafine-grained Ferrite Steel 2 . Material and Methods 3 . Results and Discussion, Mater. Res. 17 (2014) 527–534.

[70] Y. Kimura, T. Inoue, Influence of warm tempforming on microstructure and mechanical properties in an ultrahigh-strength medium-carbon low-alloy steel, Metall. Mater. Trans. A Phys. Metall. Mater. Sci. 44 (2013) 560–576.

References ________________________________________________________________________________________________________________________________

129

doi:10.1007/s11661-012-1391-2. [71] T. Inoue, F. Yin, Y. Kimura, K. Tsuzaki, S. Ochiai, Delamination effect on

impact properties of ultrafine-grained low-carbon steel processed by warm caliber rolling, Metall. Mater. Trans. A Phys. Metall. Mater. Sci. 41 (2010) 341–355. doi:10.1007/s11661-009-0093-x.

[72] A. Dolzhenko, Z. Yanushkevich, S.A. Nikulin, A. Belyakov, R. Kaibyshev, Impact toughness of an S700MC-type steel: Tempforming vs ausforming, Mater. Sci. Eng. A. 723 (2018) 259–268. doi:10.1016/j.msea.2018.03.044.

[73] T. Inoue, Y. Kimura, S. Ochiai, Shape effect of ultrafine-grained structure on static fracture toughness in low-alloy steel, Sci. Technol. Adv. Mater. 13 (2012). doi:10.1088/1468-6996/13/3/035005.

[74] Y. Kimura, T. Inoue, Y.I.N. Fuxing, K. Tsuzaki, Delamination toughening of ultrafine grain structure steels processed through tempforming at elevated temperatures, ISIJ Int. 50 (2010) 152–161. doi:10.2355/isijinternational.50.152.

[75] X. Min, Y. Kimura, T. Kimura, K. Tsuzaki, Delamination toughening assisted by phosphorus in medium-carbon low-alloy steels with ultrafine elongated grain structures, Mater. Sci. Eng. A. 649 (2016) 135–145. doi:10.1016/j.msea.2015.09.102.

[76] J. Svoboda, F.D. Fischer, P. Fratzl, E. Kozeschnik, Modelling of kinetics in multi-component multi-phase systems with spherical precipitates I: Theory, Mater. Sci. Eng. A. 385 (2004) 166–174. doi:10.1016/j.msea.2004.06.018.

[77] E. Kozeschnik, J. Svoboda, P. Fratzl, F.D. Fischer, Modelling of kinetics in multi-component multi-phase systems with spherical precipitates II: Numerical solution and application, Mater. Sci. Eng. A. 385 (2004) 157–165. doi:10.1016/j.msea.2004.06.018.

[78] H. Buken, P. Sherstnev, E. Kozeschnik, A state parameter-based model for static recrystallization interacting with precipitation, Model. Simul. Mater. Sci. Eng. 24 (2016) 35006. doi:10.1088/0965-0393/24/3/035006.

[79] H. Buken, E. Kozeschnik, A Model for Static Recrystallization with Simultaneous Precipitation and Solute Drag, Metall. Mater. Trans. A Phys. Metall. Mater. Sci. 48 (2017) 2812–2818. doi:10.1007/s11661-016-3524-5.

[80] E. Kozeschnik, A Scheil – Gulliver Model with Back- Diffusion Applied to the Microsegregation of Chromium in Fe-Cr-C Alloys, Metall. Mater. Trans. A. 31A (2000) 1682–1684.

[81] E. Kozeschnik, W. Rindler, B. Buchmayr, Scheil – Gulliver simulation with partial redistribution of fast diffusers and simultaneous solid-solid phase transformations, Int. J. Mater. Res. 98 (2007) 826–831. doi:10.3139/146.101535.

[82] J. Kreyca, E. Kozeschnik, State parameter-based constitutive modelling of stress strain curves in Al-Mg solid solutions, Int. J. Plast. 103 (2018) 67–80. doi:10.1016/j.ijplas.2018.01.001.

[83] B. Gault, M.P. Moody, J.M. Cairney, S.P. Ringer, Atom Probe Microscopy, Springer New York, New York, NY, 2012. doi:10.1007/978-1-4614-3436-8.

[84] S.A. Aksenov, Y.A. Puzino, I.P. Mazur, Inverse Analysis of Plane Strain and Uniaxial Compression Tests Performed on Gleeble, in: Metal, 2015: p. 6.

[85] Y. Xu, J. Zhang, Y. Bai, M.A. Meyers, Shear localization in dynamic deformation: Microstructural evolution, Metall. Mater. Trans. A Phys. Metall. Mater. Sci. 39 A (2008) 811–843. doi:10.1007/s11661-007-9431-z.

References ________________________________________________________________________________________________________________________________

130

[86] M. Charleux, W.J. Poole, M. Militzer, A. Deschamps, Precipitation HSLA-Nb / Ti Behavior Its Effect on Strengthening of an, Metall. Mater. Trans. A. 32 (2001) 1635–1647.

[87] J. Li, P. Xu, Y. Tomota, Y. Adachi, Dynamic Recrystallization Behavior in a Low-carbon Martensite Steel by Warm Compression, ISIJ Int. 48 (2008) 1008–1013. doi:10.2355/isijinternational.48.1008.

[88] R.D.K. Misra, H. Nathani, J.E. Hartmann, F. Siciliano, Microstructural evolution in a new 770 MPa hot rolled Nb-Ti microalloyed steel, Mater. Sci. Eng. A. 394 (2005) 339–352. doi:10.1016/j.msea.2004.11.041.

[89] ISO, ISO 18265 Metallic materials - Conversion of hardness values, 2003. [90] C. Ledermueller, H. Li, S. Primig, Engineering Hierarchical Microstructures

via Advanced Thermo-Mechanical Processing of a Modern HSLA Steel, Metall. Mater. Trans. A. 49 (2018) 6337–6350. doi:10.1007/s11661-018-4934-3.

[91] R.. Doherty, D.A. Hughes, F.J. Humphreys, J.J. Jonas, D. Juul Jensen, M.E. Kassner, W.E. King, T.R. McNelley, H.J. McQueen, A.D. Rollett, Current issues in recrystallization a review, Mater. Sci. Eng. A. 238 (1997) 219–274.

[92] R.A. Petković, M.J. Luton, J.J. Jonas, Recovery and recrystallization of carbon steel between intervals of hot working, Can. Metall. Q. 14 (1975) 137–145. doi:10.1179/000844375795050201.

[93] R.A. Grange, C.R. Hribal, L.F. Porter, Hardness of Tempered Martensite in Carbon and Low Alloy Steels, Metall. Trans. A. 8A (1977) 1775–1785. doi:10.1007/BF02646882.

[94] J.M. Rosenberg, H.R. Piehler, Calculation of the taylor factor and lattice rotations for bcc metals deforming by pencil glide, Metall. Trans. 2 (1971) 257–259. doi:10.1007/BF02662666.

[95] S.H.M. Anijdan, M. Hoseini, S. Yue, Texture development in cool deformed microalloyed steels, Mater. Sci. Eng. A. 528 (2011) 6788–6793. doi:10.1016/j.msea.2011.05.048.

[96] F.T. Han, Z.C. Wang, C.N. Jing, X.M. Liu, J. Su, S.Y. Zhang, Microstructure and Microtexture Evolution during Batch Annealing of Warm-Rolled Ti-IF Steel Sheets, Appl. Mech. Mater. 331 (2013) 443–447. doi:10.4028/www.scientific.net/AMM.331.443.

[97] M.R. Toroghinejad, A.O. Humphreys, D. Liu, F. Ashrafizadeh, A. Najafizadeh, J.J. Jonas, Effect of rolling temperature on the deformation and recrystallization textures of warm-rolled steels, Metall. Mater. Trans. A Phys. Metall. Mater. Sci. 34 A (2003) 1163–1174. doi:10.1007/s11661-003-0136-7.

[98] M.R. Barnett, J.J. Jonas, Influence of ferrite rolling temperature on grain size and texture in annealed low C and IF steel, IS. 37 (1997) 706–714.

[99] A.O. Humphreys, D. Liu, M.R. Toroghinejad, J.J. Jonas, Effect of Chromium and Boron Additions on the Warm Rolling Behavior of Low Carbon Steels, ISIJ Int. 42 (2002) S52–S56. doi:10.1179/cmq.2006.45.4.451.

[100] Z. Jia, R.D.K. Misra, R. O’Malley, S.J. Jansto, Fine-scale precipitation and mechanical properties of thin slab processed titanium-niobium bearing high strength steels, Mater. Sci. Eng. A. 528 (2011) 7077–7083. doi:10.1016/j.msea.2011.05.073.

[101] M. Kapoor, R. O’Malley, G.B. Thompson, Atom Probe Tomography Study

References ________________________________________________________________________________________________________________________________

131

of Multi-microalloyed Carbide and Carbo-Nitride Precipitates and the Precipitation Sequence in Nb-Ti HSLA Steels, Metall. Mater. Trans. A Phys. Metall. Mater. Sci. 47 (2016) 1984–1995. doi:10.1007/s11661-016-3398-6.

[102] D.P. Clausing, Effect of plastic strain state on ductility and toughness, Int. J. Fract. Mech. 6 (1970) 71–85. doi:10.1007/BF00183662.

[103] A. V. Sergueeva, J. Zhou, B.E. Meacham, D.J. Branagan, Gage length and sample size effect on measured properties during tensile testing, Mater. Sci. Eng. A. 526 (2009) 79–83. doi:10.1016/j.msea.2009.07.046.

[104] Y.H. Zhao, Y.Z. Guo, Q. Wei, T.D. Topping, A.M. Dangelewicz, Y.T. Zhu, T.G. Langdon, E.J. Lavernia, Influence of specimen dimensions and strain measurement methods on tensile stress-strain curves, Mater. Sci. Eng. A. 525 (2009) 68–77. doi:10.1016/j.msea.2009.06.031.

[105] S. Ghosh, A.K. Singh, S. Mula, P. Chanda, V. V. Mahashabde, T.K. Roy, Mechanical properties, formability and corrosion resistance of thermomechanically controlled processed Ti-Nb stabilized IF steel, Mater. Sci. Eng. A. 684 (2017) 22–36. doi:10.1016/j.msea.2016.12.034.

[106] N. Tsuji, S. Okuno, Y. Koizumi, Y. Minamino, Toughness of Ultrafine Grained Ferritic Steels Fabricated by ARB and Annealing Process, Mater. Trans. 45 (2005) 2272–2281. doi:10.2320/matertrans.45.2272.

[107] C. Ledermueller, E. Kozeschnik, R.F. Webster, S. Primig, Advanced Thermo-mechanical Process for Homogenous Hierarchical Microstructures in HSLA Steels, Metall. Mater. Trans. A Phys. Metall. Mater. Sci. 50 (2019) 5800–5815. doi:10.1007/s11661-019-05486-5.

[108] T. Sakai, A. Belyakov, R. Kaibyshev, H. Miura, J.J. Jonas, Dynamic and post-dynamic recrystallization under hot, cold and severe plastic deformation conditions, Prog. Mater. Sci. 60 (2014) 130–207. doi:10.1016/j.pmatsci.2013.09.002.

[109] I.L. Dillamore, H. Katoh, The Mechanisms of Recrystallization in Cubic Metals with Particular Reference to Their Orientation-Dependence, Met. Sci. 8 (1974) 73–83. doi:10.1179/msc.1974.8.1.73.

[110] L. Li, W. Yang, Z. Sun, Dynamic recrystallization of ferrite with particle-stimulated nucleation in a low-carbon steel, Metall. Mater. Trans. A Phys. Metall. Mater. Sci. 44 (2013) 2060–2069. doi:10.1007/s11661-012-1572-z.

[111] L. Cheng, Q. Cai, J. Lv, W. Yu, H. Miura, Superdense microbands strengthening of textured low alloy ferritic steel, J. Alloys Compd. 746 (2018) 482–489. doi:10.1016/j.jallcom.2018.02.297.

[112] F.J. Humphreys, M. Hatherly, Recrystallization and related annealing phenomena, Elsevier Ltd, 2004.

[113] A. Das, H.W. Viehrig, F. Bergner, C. Heintze, E. Altstadt, J. Hoffmann, Effect of microstructural anisotropy on fracture toughness of hot rolled 13Cr ODS steel – The role of primary and secondary cracking, J. Nucl. Mater. 491 (2017) 83–93. doi:10.1016/j.jnucmat.2017.04.059.

[114] Y. Kimura, T. Inoue, Influence of Carbon Content on Toughening in Ultrafine Elongated Grain Structure Steels, ISIJ Int. 55 (2015) 1135–1144. doi:10.2355/tetsutohagane.100.1104.

[115] H.C. Wang, C. Somsen, Y.J. Li, S.G. Fries, E. Detemple, G. Eggeler, Effect of Nb on improving the impact toughness of Mo-containing low-alloyed steels, J. Mater. Sci. 54 (2019) 7307–7321. doi:10.1007/s10853-019-03374-2.

References ________________________________________________________________________________________________________________________________

132

[116] S.-H. Song, Z.-X. Yuan, D.-D. Shen, J. Liu, L.-Q. Weng, Grain boundary segregation of phosphorus and molybdenum in a Cr – Mo low alloy steel, Mater. Sci. Technol. 20 (2004) 117–120. doi:10.1179/026708304225010307.

[117] C.Y. Chen, C.C. Chen, J.R. Yang, Microstructure characterization of nanometer carbides heterogeneous precipitation in Ti-Nb and Ti-Nb-Mo steel, Mater. Charact. 88 (2014) 69–79. doi:10.1016/j.matchar.2013.11.016.

[118] J.G. Speer, S. Mehta, S.S. Hansen, Composition of Vanadium Carbonitride Precipitates in Microallyed Austenite, Scr. Metall. 18 (1984) 14-211241–1244.

[119] M. Nöhrer, W. Mayer, S. Primig, S. Zamberger, E. Kozeschnik, H. Leitner, Influence of deformation on the precipitation behavior of Nb(CN) in austenite and ferrite, Metall. Mater. Trans. A Phys. Metall. Mater. Sci. 45 (2014) 4210–4219. doi:10.1007/s11661-014-2373-3.

[120] J.H. Jang, C.H. Lee, Y.U. Heo, D.W. Suh, Stability of (Ti, M)C (M = Nb, V, Mo and W) carbide in steels using first-principles calculations, Acta Mater. 60 (2012) 208–217. doi:10.1016/j.actamat.2011.09.051.

[121] Y. Zhou, Z. Wang, J. Zhao, Z. Leng, Z. Niu, C. Guo, Z. Zhang, Z. Yang, C. Yao, F. Jiang, Energy for the interface system of (Nb, Mo)C/γ-Fe, Appl. Phys. A Mater. Sci. Process. 123 (2017) 1–10. doi:10.1007/s00339-017-1120-1.

[122] S. Dhara, R.K.W. Marceau, K. Wood, T. Dorin, I.B. Timokhina, P.D. Hodgson, Precipitation and clustering in a Ti-Mo steel investigated using atom probe tomography and small-angle neutron scattering, Mater. Sci. Eng. A. 718 (2018) 74–86. doi:10.1016/j.msea.2018.01.070.

[123] E. V Pereloma, I.B. Timokhina, J.J. Jonas, M.K. Miller, Microstructural characterization of warm rolled Cr-containing low carbon steel, Scr. Mater. 56 (2007) 521–524. doi:10.1016/j.scriptamat.2006.11.027.

[124] K.Y. Xie, T. Zheng, J.M. Cairney, H. Kaul, J.G. Williams, F.J. Barbaro, C.R. Killmore, S.P. Ringer, Strengthening from Nb-rich clusters in a Nb-microalloyed steel, Scr. Mater. 66 (2012) 710–713. doi:10.1016/j.scriptamat.2012.01.029.

[125] S. Primig, G. Stechauner, E. Kozeschnik, Early Stages of Cu Precipitation in 15–5 PH Maraging Steel Revisited − Part I: Experimental Analysis, Steel Res. Int. 88 (2017) 1–10. doi:10.1002/srin.201600084.

[126] Z. Wang, X. Fang, H. Li, W. Liu, Atom Probe Tomographic Characterization of Nanoscale Cu-Rich Precipitates in 17-4 Precipitate Hardened Stainless Steel Tempered at Different Temperatures, Microsc. Microanal. 23 (2017) 340–349. doi:10.1017/S1431927616012629.

Appendix A ________________________________________________________________________________________________________________________________

133

10 Appendix A: MatCalc script

$$ date: 2018-07-16 Parameter study of simplified chemistry (Fe, Mn, C, N, Nb, Ti) only cementite and NbTiCN simulation of precipitates that occur during casting then stop at 950 for solution treatment and WQ Simulation of thermo-mechanical processing (heating to 600C with heating rate of 10/s, three hit deformation with strain rate of 10, and a strain of 0.2 each, aging for 60min) Script written by C Ledermueller under guidance of E. Kozeschnik $$ $ close any existing workspace without asking for save ... close-workspace f $ make sure we are in the correct working directory! set-working-directory . $ let's rock ... new-workspace open-thermodynamic-database mc_fe.tdb select-elements fe c Mn nb ti n select-phases fcc bcc_a2 cementite read-thermodynamic-database read-mobility-database mc_fe.ddb enter-composition weight-percent c=0.047 mn=1.92 nb=0.045 ti=0.016 n=0.0054 change-phase-status FCC_A1#01 major-constituents=:NB:C: create-new-phase FCC_A1 equilibrium change-phase-status FCC_A1#02 major-constituents=:TI:N: set-automatic-startvalues set-temperature-celsius 1400 calculate-equilibrium create-tm-treatment tmt $ create a new heat treatment append-tmt-segment tmt edit-tmt-segment tmt . precipitation-domain=austenite $ set precipitation domain to austenite

Appendix A ________________________________________________________________________________________________________________________________

134

edit-tmt-segment tmt . segment-start-temperature=1478 $ set start temperature of heat treatment to 1478C...from scheil-calc edit-tmt-segment tmt . T_end+T_dot 1450 0.16 append-tmt-segment tmt $ append new segment to heat treatment edit-tmt-segment tmt . T_end+T_dot 950 0.16 $ cool down append-tmt-segment tmt edit-tmt-segment tmt . T_dot+delta_t 0 600 $ 10min solution annealing append-tmt-segment tmt edit-tmt-segment tmt . T_end+T_dot 492 15 $ martensite-start temp from T0-calc append-tmt-segment tmt edit-tmt-segment tmt . precipitation-domain=ferrite edit-tmt-segment tmt . T_end+T_dot 25 15 append-tmt-segment tmt edit-tmt-segment tmt . T_end+T_dot 600 10 append-tmt-segment tmt edit-tmt-segment tmt . T_dot+delta_t 0 600 $10min soaking append-tmt-segment tmt edit-tmt-segment tmt . segment-accumulated-strain=0.2 edit-tmt-segment tmt . deformation-rate=1 append-tmt-segment tmt edit-tmt-segment tmt . T_dot+delta_t 0 30 append-tmt-segment tmt edit-tmt-segment tmt . segment-accumulated-strain=0.2 edit-tmt-segment tmt . deformation-rate=1 append-tmt-segment tmt edit-tmt-segment tmt . T_dot+delta_t 0 30 append-tmt-segment tmt edit-tmt-segment tmt . segment-accumulated-strain=0.2 edit-tmt-segment tmt . deformation-rate=1 edit-tmt-segment tmt . append-tmt-segment tmt edit-tmt-segment tmt . T_dot+delta_t 0 3600 $ 60min aging append-tmt-segment tmt edit-tmt-segment tmt . T_end+T_dot 25 0.16

Appendix A ________________________________________________________________________________________________________________________________

135

$-------------------------------------------- TABLES ----------------------------------------------$ $$************************************************************************************************** PRECIPITATION DOMAINS, PRECIPITATES **************************************************************************************************$$ $------------------------------------- PRECIPITATION DOMAINS --------------------------------------$ create-precipitation-domain austenite $ austenite is precipitation domain (matrix) set-precipitation-parameter austenite thermodynamic-matrix-phase=fcc_a1 $ matrix phase of domain (austenite) create-precipitation-domain ferrite $ ferrite is precipitation domain (matrix) set-precipitation-parameter ferrite thermodynamic-matrix-phase=bcc_a2 $ matrix phase of domain ferrite set-precipitation-parameter ferrite subgrain-evolution-model=abc set-precipitation-parameter ferrite microstructure-evolution substructure abc-parameters dislocation-generation-coefficient-C=1e-4 set-precipitation-parameter ferrite microstructure-evolution substructure abc-parameters similitude-parameter-A-prime=40 set-precipitation-parameter ferrite initial-grain-diameter=100e-6 set-precipitation-parameter ferrite initial-subgrain-diameter=100e-6 set-precipitation-parameter ferrite special-options mobilities subgrain-boundary-intrinsic-M0=1.0*MOB_LAGB$@ set-precipitation-parameter ferrite mechanical-properties solid-solution-strengthening sss-strengthening-coefficients C*1e-2 set-precipitation-parameter ferrite mechanical-properties general hall-petch-coefficient-sgb=10 $-------------------------------------- PRECIPITATE PHASES ----------------------------------------$ create-new-phase fcc_a1#01 precipitate NbC $ create new precipitate phase NbC_p0 in austenite new-workspace set-precipitation-parameter fcc_a1#01_p0 nucleation-sites=dislocations $ nucleation sites are dislocations set-precipitation-parameter FCC_A1#01_P0 interstitial-diffusion-in-precipitate-factor=1e-6 set-precipitation-parameter FCC_A1#01_P0 substitutional-diffusion-in-precipitate-factor=1e-6 set-precipitation-parameter fcc_a1#01_p0 nucleate-only-with-valid-major-constituents=yes

Appendix A ________________________________________________________________________________________________________________________________

136

set-precipitation-parameter FCC_A1#01_P0 nucleation-sites=subgrain-boundaries create-new-phase fcc_a1#02 precipitate TiN(gamma) $ create new precipitate phase TiN_p0 in austenite set-precipitation-parameter fcc_a1#02_p0 nucleation-sites=dislocations $ nucleation sites are dislocations set-precipitation-parameter FCC_A1#02_P0 restrict-nucleation-to-precipitation-domain=austenite create-new-phase fcc_a1#02 precipitate TiN(alpha) $ create new precipitate phase TiN_p0 in austenite set-precipitation-parameter fcc_a1#02_p1 nucleation-sites=dislocations $ nucleation sites are dislocations and subgrain boundaries set-precipitation-parameter FCC_A1#02_P1 interstitial-diffusion-in-precipitate-factor=1e-6 set-precipitation-parameter FCC_A1#02_P1 substitutional-diffusion-in-precipitate-factor=1e-6 set-precipitation-parameter fcc_a1#02_p1 nucleate-only-with-valid-major-constituents=yes set-precipitation-parameter FCC_A1#02_P1 nucleation-sites=subgrain-boundaries set-precipitation-parameter FCC_A1#02_P1 restrict-nucleation-to-precipitation-domain=ferrite create-new-phase cementite precipitate cem $ create new precipitate phase Q_p0 in austenite $set-precipitation-parameter cementite_p0 nucleation-sites=dislocations $ nucleation sites are dislocations set-precipitation-parameter CEMENTITE_P0 nucleation-sites=grain-boundaries $$************************************************************************************************** ********************************* OUTPUT WINDOWS, PLOTS, ETC. ************************************** **************************************************************************************************$$ new-gui-window p1 $ generate new plot: temperature $---------------- Define values for default x-axis (will be used by all plots) -----------------$ set-gui-window-property . default-x-axis-data=stepvalue $ default x-axis variable (time) set-gui-window-property . default-x-axis-for-all-plots=yes $ use default x-axis for all plots: yes set-gui-window-property . default-x-axis-title=time / s $ default x-axis title set-gui-window-property . default-x-axis-factor=1 $ scaling factor is 1 $set-gui-window-property . number-of-plot-columns=2 $ 2 plot columns

Appendix A ________________________________________________________________________________________________________________________________

137

set-plot-option . series new buffer t$c $ add series: temperature set-plot-option . series rename -1 t$c T $ define series legend set-plot-option . y-axis-title=temperature / C $ y-axis title set-plot-option . legend-alignment=none $ no legend create-new-plot xy-plot . $ create new plot: mean radii set-plot-option . use-alias-names-in-legend=yes $ replace variable names by kinetic alias set-plot-option . series new buffer f_prec$* $ add all series: phase fractions of precipitates set-plot-option . y-axis-title=phase fraction $ change y-axis title set-plot-option . y-axis-type=log $ use log scale for y-axis set-plot-option . y-axis-scaling=1e-8.. $ scale the y-axis from 1e-8.. set-plot-option . legend-alignment=bottom $ show legend at the bottom of the figure create-new-plot xy-plot . $ create new plot: mean radii set-plot-option . use-alias-names-in-legend=yes $ replace variable names by kinetic alias set-plot-option . series new buffer r_mean$* $ add all series: mean radii of precipitates set-plot-option . y-axis-title=mean radius / nm $ change y-axis title set-plot-option . y-axis-type=log $ use logarithmic scale for y-axis set-plot-option . y-axis-factor=1e9 $ scaling factor is 1e9 set-plot-option . legend-alignment=bottom $ show legend at the bottom of the figure create-new-plot xy-plot . $ create new plot: number densities set-plot-option . use-alias-names-in-legend=yes $ replace variable names by kinetic alias set-plot-option . series new buffer num_part$* $ add all series: number densities of precipitates set-plot-option . y-axis-title=number density / m<sup>-3</sup> $ change y-axis title set-plot-option . y-axis-type=log $ use logarithmic scale for y-axis set-plot-option . y-axis-scaling=1.. $ scale the y-axis from 1..

Appendix A ________________________________________________________________________________________________________________________________

138

set-plot-option . legend-alignment=bottom $ show legend at the bottom of the figure $~~~~~~~plots for dislocation density, subgrain size and yield strength~~~~~~~~ new-gui-window p1 set-gui-window-property . default-x-axis-data=stepvalue $ default x-axis variable (time) set-gui-window-property . default-x-axis-for-all-plots=yes $ use default x-axis for all plots: yes set-gui-window-property . default-x-axis-title=time / s $ default x-axis title set-gui-window-property . default-x-axis-factor=1 $ scaling factor is 1 set-plot-option . series new buffer t$c $ add series: temperature set-plot-option . series rename -1 t$c T $ define series legend set-plot-option . y-axis-title=temperature / C $ y-axis title set-plot-option . legend-alignment=none $ no legend create-new-plot xy-plot . set-plot-option . use-alias-names-in-legend=yes $ replace variable names by kinetic alias set-plot-option . series new buffer DD_TOT$ferrite $ add all series: total dislocation density set-plot-option . y-axis-title=total dislocation density / m<sup>-2</sup> $ change y-axis title set-plot-option . y-axis-type=log $ use log scale for y-axis set-plot-option . y-axis-scaling=auto $ scale the y-axis automatic set-plot-option . legend-alignment=bottom $ show legend at the bottom of the figure create-new-plot xy-plot . set-plot-option . use-alias-names-in-legend=yes $ replace variable names by kinetic alias set-plot-option . series new buffer SGD$ferrite $ add all series: subgrain diameter set-plot-option . y-axis-title=subgrain diameter / µm $ change y-axis title set-plot-option . y-axis-type=log $ use log scale for y-axis set-plot-option . y-axis-scaling=auto $ scale the y-axis automatic set-plot-option . legend-alignment=bottom $ show legend at the bottom of the figure set-plot-option . y-axis-factor=1e006

Appendix A ________________________________________________________________________________________________________________________________

139

create-new-plot xy-plot . set-plot-option . use-alias-names-in-legend=yes $ replace variable names by kinetic alias set-plot-option . series new buffer TYS_0$ferrite $ add all series: yield strength set-plot-option . series new buffer TSSS$ferrite set-plot-option . series new buffer TDS$ferrite set-plot-option . series new buffer TSGS$ferrite set-plot-option . series new buffer TSIGMA_PREC$ferrite set-plot-option . y-axis-title=yield strength / MPa $ change y-axis title set-plot-option . y-axis-type=lin $ use log scale for y-axis set-plot-option . y-axis-scaling=auto $ scale the y-axis automatic set-plot-option . legend-alignment=bottom $ show legend at the bottom of the figure set-plot-option . y-axis-factor=1e-006 $$************************************************************************************************** ************************************ SIMULATION SETUP ********************************************** **************************************************************************************************$$ set-simulation-parameter end-time=1e24 $ set simulation end time set-simulation-parameter tm-treatment-name=tmt $ T-control from defined heat treatment set-simulation-parameter max-temperature-step=10 set-simulation-parameter starting-conditions=reset-precipitates $ simulation starting conditions: reset $ speed up simulations, modify numerical parameters $set-simulation-parameter max-radius-change-during-growth=1.0 $ maximum radius growth from 0.2 to 1.0 $$************************************************************************************************** ******************************** START PRECIPITATE SIMULATION ************************************** **************************************************************************************************$$ start-precipitate-simulation $$**************************************************************************************************

Appendix A ________________________________________________________________________________________________________________________________

140

******************************* PRECIPITATE SIMULATION FINISHED ************************************ **************************************************************************************************$$

Microstructural engineering of modern high strength low alloy ...

Documents