Development of a heat treatment method to form

Development of a heat treatment method to form
Development of a heat treatment method to form
a duplex microstructure of lower bainite and
martensite in AISI 4140 steel
Erik Claesson
Master Thesis
Department of Material Science and Engineering
Royal Institute of Technology
Stockholm, Sweden
2014
By: Erik Claesson
Degree project report at
Degree Program in Materials Design and Engineering
Royal Institute of Technology, KTH, Stockholm
Supervisor: Lisa Earnest
Cummins Fuel Systems
Cummins Inc, 1460 N National Rd, Columbus, IN 47201
Examiner: Prof. Joakim Odqvist
Department of Metallography, School of Industrial Engineering and
Management, Royal Institute of Technology, KTH, Stockholm
June 2014
i
Abstract
Research on bainite and martensite structures has indicated that lower bainite needles have a
refining effect on the lath martensitic structure. Lower bainte needles partitions prior austenite
grains and will consequently have a refining effect on the subsequent formed lath martensite.
Smaller austenite grains will result in smaller lath martensitic packets and blocks and will result
in enhanced mechanical properties.
In order to create a variation of lower bainte structure in a matrix of martensite, two different
heat treating methods were tested. The work was focused towards the formation of lower bainite
during isothermal heat treating in molten salt, above and below the MS-temperature. Both untempered and tempered samples were analyzed .Two different materials were tested, both were
AISI 4140 but with a slightly difference in hardenability. The material provided by Ovako Steel
is 326C and 326F the later had a higher hardenability.
In order to better distinguish the two structures from each other when studied under a
microscope, a variation of etching methods were tested.
It was possible to create a variation of lower bainite structures in a matrix of martensite. 326F
shows less amount of lower bainite and provides a higher average surface hardness before
tempering.
Keywords: AISI 4140, 326C, 326F, Isothermal heat treatment, Martensite, Bainite, Molten salt,
vacuum furnace, gas quenching.
ii
Sammanfattning
Forsking på mikrostrukturer med en kombination av bainit och martensit har visat att underbainit
har en förfinande effekt på den martensitiska strukturen. De underbainitiska nålarna delar upp de
austenitiskakornen, vilket leder till den latt-martensitiska understrukturen blir finare. Den
förfinande strukturen resulterar i högre mekaniska egenskaper.
För att skapa olika halter av underbaint in en matris av martensit, testades olika
värmebehandlingar.
Arbetet
fokuserades
mot
underbaintisk
formation
under
isotermiskvärmebehandling i smalt salt. Två olika temperaturer av isotermiskvärmebehandling
testades, över och under MS. Även med eller utan ett avslutande anlöpnings steg testades. Två
olika material undersöktes bada AISI 4140 men med skillnad hårdbarhet. Materialen från Ovako
Steel AB heter 326C och 326F, den senare hade en högre hårdbarhet.
For att skilja martensit och bainte från varandra när de studeras under ett mikroskop testades
också olika etsnings metoder for att bättre kunna skilja underbaint och martesnite från varandra.
Det var möjligt att skapa olika halter av underbaint i martensit. Material 326F förenklade bainit
formationen samt att den gav en högre genomsnittlig hårdhet före anlöpning.
Nyckelord: AISI 4140, 326C, 326F, Isotermisk värmebehandling, Martensit, Bainit, salt,
vakuum ugn, gas släckning
iii
iv
Acknowledgement
This work has been a co-operation between Cummins Fuel Systems and the Royal Institute of
technology (KTH) in Stockholm, but also a part of my time at Purdue University. The work was
conducted at the Cummins Fuel Systems Plant in Columbus Indiana, during the spring of 2014.
First of all I would like to thank everybody at the Material Engineering department at Cummins
Fuel Systems in Columbus Indian, USA, for this great opportunity.
A big thanks to Lisa Earnest who helped me to get a better understanding about the fuels systems
business and who gladly answered all my questions. I would like to show great appreciation to
Madeline Fogler for the feedback and guidance during the whole project and for the general help
with the writhing of the thesis.
Charles Thomas helped me with the practical work and with his great knowledge about heat
treating processes provided me with a lot of interesting and necessary information to bring this
work forward, for which I am grateful.
Thanks to Terry Parsons and the technicians at the material engineering department at Cummins
Fuel systems for helping me with practical work, analyzing the samples.
I would like to thank Dr. Joakim Odquist who was my supervisor at the KTH for his guidance
and the feedback he provided during the whole period.
Erik Claesson
Columbus Indiana, USA, May, 2014
v
Nomenclature
MS
Martensitic start temperature
MF
Martensitic finish temperature
FCC
Face central cubic
BCC
Body central cubic
BCT
Body central tetragonal
σ
Tensile strength
LOM
Light optical Microscope
SEM
Secondary electron Microscopy
XRD
X-ray diffraction
σy
Total strength of lath martensite
σ0
Friction stress for pure iron
σp
Precipitation hardening
σs
Solidification hardening
σρ
dislocation hardening
KHPd1/2
Grain boundary strength, hall petch
WLB
Width of Lower Bainite Needle
LLB
Length of Lower Bainite Needle
SUM
Martensite size
SUB
Upper Bainite size
𝜎 .
Yield strength of mixed structure
𝜎 .
Yield strength Martensite
𝜎 .
Yield strength Bainite
VB
Bainite Volume
SM
Martensite Packet size
vi
k
Constant
𝜎
Friction stress
Ta
Arrest temperature, for marentsite transformation
EBSD
Electron backscattering diffraction
vii
Table of Contents
1.
Introduction............................................................................................................................ 1
1.1.
2.
Aim of the present work ................................................................................................... 2
Background ........................................................................................................................... 3
2.1. Martensite ....................................................................................................................... 3
2.1.1. Kinetics of martensite formation .......................................................................... 4
2.2. Formation of Bainite ...................................................................................................... 7
2.2.1. Diffusionless growth ............................................................................................... 8
2.2.2. Diffusional growth ................................................................................................... 9
2.3. Upper and lower bainite.............................................................................................. 10
2.4. Combination of Martensite and Bainite .................................................................... 13
2.5. Banded structures ....................................................................................................... 18
2.6. Heat treating ................................................................................................................. 21
2.6.1. Quenching ............................................................................................................. 22
2.6.2. Oil quenching ........................................................................................................ 23
2.6.3. Gas quenching...................................................................................................... 23
2.6.4. Molten salt quenching.......................................................................................... 24
2.6.5.
Isothermal heat treating temperature ................................................................... 25
2.7. AISI 4140 ...................................................................................................................... 25
2.8. Etching .......................................................................................................................... 26
2.9. Point Counting Method ............................................................................................... 27
3.
Materials .............................................................................................................................. 28
4.
Experimental procedures .................................................................................................. 30
4.1. Sample preparation ..................................................................................................... 30
4.2. Sample analysis ........................................................................................................... 30
4.3. Heat treatment ............................................................................................................. 32
4.4. Etching methods ...................................................................................................... 33
5.
Result and Discussion ....................................................................................................... 34
5.1. Vacuum heating gas quenching (VHGQ) ................................................................ 34
5.1.1. VHGQ test 1 .......................................................................................................... 34
5.2. Austempering in molten salt ...................................................................................... 36
viii
5.2.1. Trial A ..................................................................................................................... 36
5.2.2. Isothermal heat treating temperature ................................................................ 40
5.2.3. Trial B ..................................................................................................................... 41
5.2.4. Comparison of data.............................................................................................. 44
5.3. Etching methods .......................................................................................................... 48
6.
Conclusions ......................................................................................................................... 52
7.
Further work ........................................................................................................................ 52
References .................................................................................................................................. 53
ix
1.
Introduction
In the fuel system industry the production of materials with high mechanical properties i.e.
strength and hardness, is crucial in order to provide the customers with reliable products. The
everyday work at the Materials department at Cummins Fuel Systems involves solving problems
caused by heat treatment issues, rust, manufacturing defects or steel making defects. Everything
from cracks caused by inclusions to failures due to manufacturing problems are being analyzed,
and solved.
The fuel systems designed and manufactured in Columbus, Indiana consist of 6 or 8 fuel
injectors, a fuel pump and a rail connecting the pump and the injectors. The highly calibrated and
sensitive technology relies on material made to withstand high cyclic pressure for one million
miles on engine. A primary component in the injector is the injector body made from AISI 4140
(Ovako 326C) steel provided from Ovako AB. Research in order to understand the different
parameters i.e. heat treatment, machining and finishing operations is important. Knowledge
about materials and the operating environments will aid the development of new materials with
better mechanical properties. Knowledge to further understand the impact of the microstructural
result of steel heat treatment is something that has been investigated for a long time, with a high
focus on the austenite transformation and its products. The formation and distribution of
microstructures in steels will very much influence the limitations of the product.
The transformation of austenite into non-martensitic structures is often related to insufficient heat
treatment. The heat treating process has been developed to ensure no non-martensitic
transformation products and thereby desired properties. Problems have occurred due to the
difficulties to control heat treating to produce 100% martensite structure. However, it has been
shown that non-martensitic transformation into lower bainite can increase the mechanical
properties. Lower bainite is a fine and complex structure of ferrite and cementite, and have
shown to increase toughness, strength and fatigue properties.
For many years, AlSI 4140 has been heat treated to form a martensitic structure. The
characteristics of tempered martensite have been favorable when it comes to high demanding
applications. Tempered martensite has been used for its excellent mechanical properties in
applications where it is important to withstand high pressure fatigue environments. However,
problems involving distortion and micro cracking in the structure can lower the fatigue strength.
A duplex structure of martensite and lower bainite has shown to increase the tensile strength. [1]
Indications that the fatigue strength also can be increased with a combination of martensite and
lower bainite has raised the motivation to further research this phenomenon. [2] [3] [4] [5]
1
1.1.
Aim of the present work
This work is considered a pre-study before mechanical testing. Cummins and Ovako are teaming
up to further investigate AISI 4140 with mixed levels of bainite and martensite.
The main objective is to heat treat AISI 4140 in order to form different levels of bainite, measure
the bainite content by using a point counting method, and use a variation of etchants to better
distinguish bainite and martensite.
Two different materials with slightly different hardenability levels will be shipped to a heat treat
supplier in order perform isothermal heat treatment in a molten salt bath. This will form a
variation of bainite levels in a martensite matrix.
The vacuum harden and gas quench furnace at Cummins Fuel Systems in Columbus, Indiana
will be used in order to try to adjust the cooling rate to form bainite. Parameters such as pressure,
with or without use of a cooling fan, section size and variation in atmosphere will decide the
resulting microstructure.
2
2.
2.1.
Background
Martensite
Many materials form a martensitic structure by quenching. However, the diffusionless formation
of martensite is most associated with steel processing and rapid transformation of austenite. The
formation of martensite is highly influenced by the carbon content, prior austenite grain size and
the alloy content. All three parameters are also associated with steel hardenability.
The diffusionless transformation from austenite and the high strength of martensite are related to
trapped carbon atoms in the martensite lattice. Since diffusion does not have time to occur, the
chemical composition and distribution of alloying elements in the formed martensite is
determined by the converted austenite. The maximum carbon content in austenite is around 2%,
since martensite only can be formed via austenite transformation, 2% is also the maximum
carbon content in martensite. The austenite face centered cubic (FCC) lattice will upon
quenching transform into martensite body centered tetragonal (BCT) lattice via a shear
mechanism, resulting in trapped carbon atoms. The shear mechanism affects the lattice and
causes the surface to tilt, which results in plastically deformed regions close to the tilted surface
as shown in Figure 1. [6]
Figure 1. Schemeatic picture of the martensitc shear mechanism reuslting in a tilted martensite surface [6]
The habit plane is the preferred austenite plane from which the martensite crystals are nucleated.
The interface between the two phases is ideally planar, but varies with composition. The plane
where the formation is initiated is called the midrib. The midrib is a characteristic feature for the
martensitic phase, when studied under a microscope the midrib can be seen as a line in the center
of the martensitic phase dividing it in half. The formation of martensite is due to a displacive
formation. Instead of diffusion in order to adapt to the new lower energy state at lower
temperature, the carbon atoms are moved simultaneously by the shear mechanism, in opposite
3
direction on each side of the midrib plane. The forced displacement of carbon atoms will result
in high amount of internal stresses and an increase in dislocation density, resulting in a hard but
brittle material. [6]
In contrast to the ferrite body centered cubic (BCC) lattice, the martensitic BCT lattice has one
side that is longer than the other two sides, resulting in more interstitial sites, and a higher
solubility of carbon.
Martensite is a metastable phase that will decompose into ferrite and carbides (depending on
carbon content) during tempering. The decomposition of martensite during tempering can be
divided into the following steps:
1. Formation of carbides, resulting in decreased carbon content in the martensitic structure,
which will decrease the residual stresses. The carbides are a so-called transition carbides
named epsilon
2. If any retained austenite is present it will decompose into ferrite and carbides
3. Decomposition of martensite into ferrite and carbides
The time necessary for decomposition depends on the temperature, and the mobility of the
carbon atoms. [6]
2.1.1. Kinetics of martensite formation
The diffusionless transformation is an athermal transformation because it is not associated with
any thermal activation. When cooled below the martensitic start temperature (MS), the formation
of martensite will take place. Theoretically, if the temperature is cooled down and kept at a
certain temperature under MS but above the martensitic finish temperature (MF), for a long time,
no further formation of martensite will take place. MF is the temperature when the last
transformation of austenite into martensite occurs. The MS and MF are related to the amount of
driving force for martensite transformation, which is the energy needed for the shear mechanism
to take place. The MS depends on the carbon content in the steel. Alloys with higher carbon
content will have a lower corresponding MS. Hence, higher carbon content provides higher
resistance against transformation into martensite. The amount of alloying elements will also have
a decreasing effect on the MS and MF. In general, alloying elements will lower the MF with the
exception of cobalt. Different equations have been established over the years to describe the
impact of alloying elements on MS and MF, and are further described by George Krauss. [6]
When the carbon content reaches 0.3wt% and above, MF will go below room temperature,
increasing the risk for retained austenite. Retained austenite impairs the mechanical properties
4
and is not desired in Fuel Systems products. In order to ensure that no austenite will remain in
the steel after quenching, deep freezing is commonly used i.e. in liquid nitrogen. [6]
The carbon content also decides what type of martensite that will be formed. Two variations of
martensite exist, lath and plate martensite. An increase in carbon content will lower the MS and
MF temperatures. However, increasing the carbon content will consequently lead to a difference
in the appearance of martensite, as shown in Figure 2. Lath martensite is formed at lower carbon
levels while plate martensite is formed at higher levels. Lath and plate martensite are threedimensional individual crystals of martensite. When etched and studied under a microscope, lath
and plate martensite can be seen as needles or acicular plates respectively. However, the
microstructure is often too fine to see any of the needles or the acicular crystals in light optical
microscope. Bainite has a similar appearance and is one reason that these two structures can be
difficult to distinguish in a light optical microscope. [6]
Figure 2. MS-temperature ploted aginst the carbon contetn, as-well as a description of the relationship between carbon
content the apperence of lath and plate martensite [7]
Lath martensite appears in packets. A packet consists of laths with the same habit plane but with
different crystallographic orientations and can be seen as individual grains. The packets form a
substructure inside the prior austenite grain, and are an important factor when it comes to
mechanical properties. Inside each packet, laths with the same crystallographic orientation
subdivide into blocks, as seen in Figure 3 . Plate martensite does not appear in packets, which
make the formation of lath packets an important factor in low and middle range carbon steels.
[6][8]
5
Figure 3. An illustrative image of the reltionship between the prior austenite grain and the martensite packets and blocks
[9]
The strength of lath martensite can be explained by the following equation
𝝈𝒚 = 𝝈𝟎 + 𝝈𝒑 + 𝝈𝒔 + 𝝈𝝆 + 𝒌𝑯𝑷 𝒅
𝟏/𝟐
( 1)
𝜎 = 𝑓𝑟𝑖𝑐𝑡𝑖𝑜𝑛 𝑠𝑡𝑟𝑒𝑠𝑠 𝑓𝑜𝑟 𝑝𝑢𝑟𝑒 𝑖𝑟𝑜𝑛, 𝜎 = 𝑝𝑟𝑒𝑐𝑖𝑝𝑖𝑡𝑎𝑡𝑖𝑜𝑛 ℎ𝑎𝑟𝑑𝑒𝑛𝑖𝑛𝑔, 𝜎 = 𝑠𝑜𝑙𝑖𝑑𝑠𝑜𝑙𝑢𝑡𝑖𝑜𝑛 ℎ𝑎𝑟𝑑𝑒𝑛𝑖𝑛𝑔, 𝜎 = 𝑑𝑖𝑠𝑙𝑜𝑐𝑎𝑡𝑖𝑜𝑛 ℎ𝑎𝑟𝑑𝑒𝑛𝑖𝑛𝑔 𝑎𝑛𝑑 𝑘
𝑑
/
= 𝑔𝑟𝑎𝑖𝑛 𝑏𝑜𝑢𝑛𝑑𝑎𝑟𝑦 𝑠𝑡𝑟𝑒𝑛𝑔𝑡ℎ𝑒𝑛𝑖𝑛𝑔, ℎ𝑎𝑙𝑙 𝑝𝑒𝑎𝑡𝑐ℎ The strength related to the packet and the block size (d) is the Hall Petch term, and is valid for
high-angle boundaries. High angel boundaries is boundaries with a high misorentation (>15°).
[10] Both packets and blocks form high-angle boundaries and will act as barriers for moving
dislocations during plastic deformation. Smaller grains provide smaller packets which
consequently give smaller blocks. According to the Hall Petch term, smaller units (grains,
packets and laths) will increase the strength of the material. Smaller units results in more highangle boundaries and thereby more obstacles for moving dislocations. [8]
Increase in grain size will increase size of the martensite plates consequently resulting in a higher
risk for formation of microcracks. Moreover, longer and thinner martensite plates are more likely
to form micro cracks causing a decrease in toughness and strength. Microcracks forms on the tip
of the growing plate, when the strain is high enough to induce a crack. [11] Steels with high
carbon content are more prone to microcracking, which involves formation of plate martensite.
However, microcracking has also been seen in lath martensite. [7]
6
Prior austenite grain size will have a large impact on the strength of the final material. The
formation of lath marteniste in low and middle range carbon steels will help to subdivide the
grains, resulting in more obstacles for moving dislocations during plastic deformation. Lower
carbon content will lead to increased packet and block size. Finer austenite grains will
consequently result in smaller packets and thinner blocks. [10]
However, it has been shown that formation of lower bainite needles upon cooling will help to
subdivide the austenite grains further, and could therefore increase the strength even more, after
subsequent martensitic formation. The interaction between martenisite and lower baintite is
interesting, and the two structures have shown to support each other.
2.2.
Formation of Bainite
Depending on the cooling rate and alloy composition, austenite can transform in to different
phases and structures. The formation of martensite is related to rapid cooling, or diffusionless
transformation, as discussed in Section 2.1. Martensite is a hard and brittle phase, but can be
tempered to have a high toughness due to reduced residual stresses and a decrease in dislocation
density.
Slower cooling rates often results in pearlite, a lamellar structure of ferrite and cementite, which
is formed at relatively high temperatures. Ferrite and cementite grow side by side to form
pearlite, by a so-called edge growth. Pearlite is a soft and ductile structure and is not desired in
high pressure applications. However, at lower temperatures (550 C down to Ms), the
mechanisms behind the formation of ferrite and cementite are different, resulting in a nonlamellar structure. The diffusion of carbon gets sluggish at lower temperatures, resulting in a
fine and complex structure called bainite. During the transformation into bainite, ferrite is the
leading growing phase, and the formation of cementite will occur when the carbon concentration,
of the surrounding austenite or the ferrite phase, is high enough to allow precipitation of
carbides.
The growth of bainite and the mechanisms behind it have been a topic of debate for a long time.
Two main theories are generally discussed: diffusionless and diffusional theory. As the names
reveal, the theories explain independent mechanisms that depend on diffusion or diffusionless
transformation. [12]
7
2.2.1. Diffusionless growth
The diffusionless formation is described to be caused by a shear mechanism. Atoms are moving
simultaneously in a glide-type motion which is called a displacive motion. [13] The diffusionless
theory suggests that the transformation of bainite is a result of small subunits of ferrite that are
formed similar to martensite. The austenite surrounding the ferrite will have an increase in
carbon content and the formation of cementite will eventually take place. When the formation of
cementite occurs, it will result in new sites for ferrite nucleation. [12] Moreover, studies have
shown that no redistribution of substituional atoms occur during formation of bainite. Therefore,
it is possible to reject any growth mechanisms dependent on substitutional diffusion. [14]
The formation of bainite is associated with a surface relief, the same thing happens during
martensite formation. Experiments where austenite grains have been polished in order to study
the bainite formation have revealed that the surface is provided with a curvature after austenite
transformation into bainite, also called plastic relaxation. This type of phenomena is often
associated with displacive growth, and a similar phenomenon was discussed for martensite
formation in section 2.1. [14]
The decomposition of austenite into bainite is divided into the following steps:
1.
2.
3.
4.
Nucleation of plates/subunits of bainitic ferrite on the austenite grain boundary.
The next nucleated subunit is nucleated on the tip of the previous formed subunit.
A number of subunits will eventually form a so-called sheaf.
Carbide formation on the ferrite/austenite interphase, or inside the ferrite plate
The precipitation of carbides influences the reaction rate by reducing the carbon content in the
austenite or the supersaturated ferrite, as shown in Figure 4. The controlling growth mechanism
is the growth of the sheaf, which is slower than the growth of a subunit. The formation of the
later is considered martensitic. [15]
Figure 4. The nucleation of bainitic subunits takes place at the grain boundaries. Multible subunits will eventually form
a bainitc sheaf. [16]
8
2.2.2. Diffusional growth
The Diffusional theory describes the growth of bainite as a process controlled by redistribution
of atoms, which involve diffusion of individual atoms away from the ferrite/austenite interface.
[13] The diffusional theory also claims that the formation of bainite is initiated by formation of
acicular ferrite or widmanstätten ferrite plates on the austenite grain boundaries parallel to each
other. [15] In the temperature region between 550 C and MS, the driving force for formation of
ferrite is high. As the temperature decreases the equilibrium content of carbon in austenite will
increase, this can be seen studying the Fe-C diagram, by following the equilibrium line for ferrite
and eutectoid structure where the carbon content in ferrite decreases with temperature, resulting
in a redistribution of carbon to the untransformed austenite. The result is a rapid transformation
to ferrite. Between the ferrite plates a mix of ferrite and cementite is formed.
The difference between pearlite and bainite growth, according to the diffusional theory, can be
seen in Figure 5. The redistribution of carbon is necessary in order to form ferrite and cementite.
Lamellar pearlite is formed because the carbon atoms only need to move a short distance. Higher
mobility will allow carbon atoms to move longer distances which results in thicker lamellas.
However, the transformation into bainite is due to the limited diffusion rate of carbon due to the
low temperature, resulting in a finer distribution of ferrite and cementite. [12]
Diffusion of carbon increases the carbon content in austenite, resulting in carbides. The diffusion
allows low carbon ferrite to grow, and is affected by the equilibrium carbon concentration at the
interface between ferrite and austenite. [14]
As discussed for the diffusionless theory the formation of bainitic ferrite is associated with a
surface relief, and was brought up as evidence that the formation type is martensitic. However, in
the diffusional theory the formation of acicular ferrite is assumed to be identical to the formation
of widmanstätten ferrite. [12] Widmanstätten ferrite together with widmantätten cementite is also
formed with a surface relief. However, today the formation of widmanstätten ferrite and
widmantätten cementite is generally considered among researchers to be diffusional. [12]
Many scholars still consider the diffusionless theory, explaining that bainite is formed at
temperatures below pearlite transformation. However, Hillert rejects the theory for diffusionless
growth of martensite by assuming that the plate formation is the same as for acicular ferrite, and
that the acicular transformation is the same as for widmanstatten ferrite. [14]
9
Figure 5. Difference in ferrite and cementite formation in pearlite and bainite can be seperated by the distibution of the
carbides and ferrite. Pearlite will have a lamellar strucutre, which is not the case for the bainte strucutre. [12]
2.3.
Upper and lower bainite
Depending on the transformation temperature and the carbon content of the steel alloy, can result
in different kinds of bainite. Two main types of bainite are going to be discussed, upper and
lower bainite, which form at upper or lower temperatures. [12] The variation in morphologies
between the two types of bainite is due to different rates of carbon rejection, which is the
transportation of carbon away from the area, which will in a subsequent scenario transform into
ferrite. The variation in morphology between upper and lower bainite is also due to the
difference in growth mechanisms, and is related to the temperature at which the two structures
are formed.
The coherency of the interface between ferrite and austenite during growth, will affect the bainite
morphology, resulting in different ferrite appearance. Upper bainite have so called lath formed
ferrite while lower bainite consists of ferrite with plate like appearance. Upper and lower bainite
have low and high coherency respectively, resulting in a more rapid growth of the ferrite phase
in lower bainite formation. [17]
The transition temperature, where upper bainite formation shift and lower bainte starts to form, is
dependent on the carbon content, see Figure 6. The upper to lower bainite transition temperature
is explained in the work of Ławrynowicz, and is defined as, “the highest temperature at which time required to obtain chosen volume fraction of cementite precipitation is smaller than the time
necessary to decarburize ferrite lath”. [18]
10
Figure 6. The relationship between the upper to lower bainte transition temperature depends on the carbon content.[19]
During the formation of upper bainite, the carbon mobility is high and carbides are formed on the
interface between the austenite and ferrite lath. During the formation of lower bainite, the
diffusion of carbon is limited and carbides are formed inside the ferrite phase, see Figure 7. [12]
It has to be mentioned that the formed carbides do not have to be cementite; evidence of other
types of carbides formed in lower bainite has been reported. [14]
The two different bainite structures have different ferrite morphology and carbide distribution,
resulting in difference of the mechanical properties. [20] One explanation to the superior
mechanical properties of lower bainite is that lower bainite have finer and more evenly
distributed carbides, resulting in a higher necessary force in order for dislocations to pass
carbides during plastic deformation. [21] [12]
11
Figure 7. Scematic picture of the distributed carbides in upper and lower bainite. Lower bainite have carbides with
in the ferrite subunits and betwwen the ferrite plates. Upper bainte only have carbides precipitated between the
ferrite plates. [12]
The typical appearance of bainite is small plates/laths of ferrite with similar orientation, adopted
from the prior austenite grain, together forming a sheaf. [12] Lower bainite forms a needle-like
structure with cementite precipitating on the austenite/ferrite interface as well as inside the ferrite
plates. Upper bainite forms a more feathery-like structure with cementite only nucleating on the
austenite/ferrite interface. Instead of plates, the upper bainitic ferrite forms long laths. [12] From
the work of Tomita and Okabayashi, Figure 8 shows the difference in appearance of lower and
upper bainite in a matrix of martensite. [1]
a)
b)
Figure 8. Not only the carbide distribution distingusih the upper and lower bainite strcutures from each other. The feerite
morphology is also different for the two strucutres. Lower bainite have a needle like apperance and the upper bainte have
a more feathery like strucutre. a) Lower bainite in a matrix of martensite. b) Upper banite in a matrix of martensite [1]
12
2.4.
Combination of Martensite and Bainite
In alloys such as AISI 4140 and 4340, it has been shown that a mixture of lower bainite and
martensite is preferred in order to enhance the mechanical properties.
One important parameter influencing the strength of the material is the prior austenite grain size.
It has a big influence on the final properties. Smaller grains will increase strength since the size
of the martensitic lath packets and blocks is directly related to the size of the austenite grains.
[6][8]
Moreover, research has indicated that the size, shape, distribution and type of bainitic structure
will have a big impact on the martensite strength. For example, when the acicular bainite is
formed, it will subdivide regions of austenite and has a refining effect on the austenite grain size,
resulting in smaller units.
Different research reports that lower bainite in ultrahigh tensile strength steel and in the presence
of martensite will grow from the grain boundaries and take an acicular form. This particular
structure will aid the material when it comes to toughness and strength. [8] [1] How large the
refinement is of the martensitic structure, depends on the width (WLB) and length (LLB) of the
bainitic sheaf, which will affect the martensitic size (SUM). [21]
The bainitic structure forms before martensite and is therefore a controlling factor for martensite
transformation, or the distribution of martensitic phases. Finer austenite grains and acicular
bainite will together have a refining effect on the martensite packet size. Moreover, the strength
of the bainite structure will increase due to the surrounding martensite. The martensitic phase
will protect the bainite structure and limit its deformation, resulting in an increase in total
strength of the material. [2] [16] For example, ultrafine duplex microstructures have been shown
to be beneficial due to the secondary ductile phase (bainite) being alternated with a harder phase
(martensite), resulting in an increase in mechanical properties. [1]
It has also been found that there is a large difference in mechanical properties depending on
which type of bainite that is formed. Lower bainite has higher strength and ductility than upper
bainite, and when combined with martensite, the difference is even larger. [20]
The presence of upper bainite has impairing effects on the properties due to the morphology.
Instead of dividing the prior austenite grain into smaller subunits, like the lower bainite needles
does, the upper bainite structure fills up the prior austenite grain, which restricts the refinement
of the martensite structure, as seen in Figure 9. The decrease in strength of the mixed upper
bainite/martensite structure is caused by a non-uniform strain that occurs during the initial stages
of plastic deformation between upper bainite and martensite. The non-uniform strain occurs due
to high local internal stresses caused by different deformation of the two structures and initiates
close to the martensite/bainite interface. Upper bainite is deformed before martensite and will
13
therefore have an impairing effect the mechanical properties. The two structures do not support
each other, unlike what occurs in lower bainite and martensite mixed microstructure. [20]
Figure 9. Schematic picture of the partioning of the austenite grain caused by the upper and lower bainite structre. Lower
bainite have a more refining effect on the austenite grain. Better refining can to some degree increase the strength of the
martensite structure. [21]
The level of mixture and the quantity of each phase seems to influence the total strength of the
steel a great deal, and can be related to the role of mixture. Equation 2 has been developed from
the equation for role of mixture in order to describe the relationship between the strength and
bainite content. The refinements of the martensite packets have been taken into consideration.
[20]
𝟏
𝑴
𝑩
𝟐
𝝈𝒎𝒊𝒙
𝟎,𝟐 = (𝝈𝟎,𝟐 + 𝒌𝑺𝑴 )(𝟏 − 𝑽𝑩 ) + 𝝈𝟎,𝟐 𝑽𝑩
𝜎
= 𝑌𝑖𝑒𝑙𝑑 𝑠𝑡𝑟𝑒𝑛𝑔𝑡ℎ 𝑜𝑓 𝑚𝑖𝑥𝑒𝑑 𝑠𝑡𝑢𝑐𝑡𝑢𝑟𝑒
,
𝜎
,
= 𝑌𝑖𝑒𝑙𝑑 𝑠𝑡𝑟𝑒𝑛𝑔𝑡ℎ 𝑜𝑓 𝑚𝑎𝑟𝑡𝑒𝑛𝑠𝑖𝑡𝑒
𝜎
,
= 𝑌𝑖𝑒𝑙𝑑 𝑠𝑡𝑟𝑒𝑛𝑔𝑡ℎ 𝑜𝑓 𝑏𝑎𝑖𝑛𝑖𝑡𝑒
𝑉 = 𝑉𝑜𝑙𝑢𝑚𝑒 𝑏𝑎𝑖𝑛𝑖𝑡𝑒
𝑆 = 𝑀𝑎𝑟𝑒𝑛𝑠𝑖𝑡𝑒 𝑝𝑎𝑐𝑘𝑒𝑡 𝑠𝑖𝑧𝑒
𝑘 = 𝐶𝑜𝑛𝑠𝑡𝑎𝑛𝑡
14
( 2)
The equation is relied upon two assumptions
The average lath width and packet size, SM, will decrease when the bainite level, VB,
increases. This results in a refinement of the substructure, and is responsible for the
increase in strength related to the Hall Petch relation, see equation 3.
𝝈𝑴
𝟎,𝟐
𝟏
𝟐
= 𝝈𝒊 + 𝒌𝑺𝑴 ( 3)
𝜎 = 𝐹𝑟𝑖𝑐𝑡𝑖𝑜𝑛 𝑠𝑡𝑟𝑒𝑠𝑠
The second assumption is that 𝜎 . , for bainite, 𝜎
fully bainitic steel. [20]
in equation1 is approximated to be
.
Lower bainite contents up to 25%, will result in substantial improvements of the mechanical
properties, and caused by the high deformation strain of the bainite structure. The 𝜎 . reaches
that of 𝜎 . and will result in a new expression for 𝜎 . , see equation 4.
𝟏/𝟐
𝝈𝒎𝒊𝒙
𝟎,𝟐 = 𝝈𝒊 + 𝒌𝑺𝑴
( 4)
Equation 4 is valid up 25% lower bainite. Above 25%, equation 2 correlates well with
experimental data. [20]
It has been demonstrated that the ferrite plates in the bainite structure can change a cracks path
during its advancement in the material, thereby delay or even prevent cracks from propagating.
[4] This phenomenon, together with an increase in strength, has shown to be valid only at lower
levels of bainite, and its contribution will decrease at higher bainite content, see Figure 10. [22]
Studies also show that for a given prior austenite grain size, the martensite packet size will
decrease with higher levels of bainite. [2]
15
Figure 10. AISI 4140 steel, strength as a function of lower bainite content. There is a peak in strength at 25% lower
bainite. [1]
Even though grain size is not a varying parameter in this study, it is important to have it in mind,
in order to understand the transformation process and the resulting microstructure and the
influences it will have on the final material. In order to control the microstructure, grain size etc.,
it is necessary to understand the heat treatment influencing the transformation of austenite, which
is determined by carbon and alloy content, cooling rate and stresses in the material. [23]
The prior austenite grain size do not only influence the martensitic packet/blocket size and
bainite sheaf size, it will also affect the transformation kinetics of bainite during its formation.
Lambers et.al studied the effect of austenitzation before transformation into bainite. They
concluded that that the larger austenite grains will increase the incubation time for bainite
formation. [24] They also concluded that the presence of residual carbides after austenization
would lower the incubation time, due to more nucleation sites. [24]
Bhadeshia claims in his report on the bainite structure, that the difference between upper and
lower bainite is the carbide distribution. [15] The formation and appearance of the lower bainite
structure can vary, depending on temperature. The hardness of the bainite structure depends on
the distribution of carbides, which also helps to distinguish between upper and lower bainite.
[15]
16
Ohtani et.al claims that different types of upper bainite can form in low carbon steels. Depending
on the temperature the carbide distribution together with the morphology of the upper bainite
sheaf will vary. They say that three different kinds of upper bainite exist. The difference between
the types is the distribution and the shape of the carbides, and the morphology of the ferrite
phase. One specific type called upper bainite III is described as having cementite particles within
the ferrite plate, resulting in better strength. They say that it is reasonable to classify this
structure as a type of upper bainite because it has typical upper bainite ferrite morphology, in
form of laths. The work of Ohtani et.al further illustrates the disagreement among researchers
about the definition and complexity of the bainite structure. [25]
During continuous cooling, the formation of needles and sheaves that are associated with bainitic
structure is dependent on the temperature during cooling. During the formation of the lower
bainite structure, the carbide distribution and morphology of the bainitic ferrite are also
dependent on temperature. [1][15] Quenching to lower temperatures is associated with a high
driving force for transformation, but the lower temperature restricts the carbon diffusion. Lower
carbon diffusion and high driving force for ferrite formation will lead to the formation of the
lower bainitic structure and the formation of long needles. The carbon atoms do not have enough
mobility to diffuse across longer distances and therefore carbides are formed inside the ferrite
subunits, creating the backbone of the lower bainitic sheaf /needle. [3]
Higher driving forces lead to more nucleation sites; together with the low diffusion and smaller
grains the number of nucleation sites will increase. At low temperatures the growth of the lower
bainitic needle and the amount of formed needles will lead to increased refinement of the
austenite grain. The strengthening of the bainitic structure is due to the fact that the stronger
surrounding martensite will restrict the deformation, enhancing the strength. After a certain
amount of grain refining and up to 25% lower bainite, the strengthening mechanism will no
longer have an increasing factor due to lower bainite having to carry the majority of the load. [1]
In order to get thin bainitic needles the transformation temperature seems to be the controlling
factor. Longer time isothermally heat treated does not seem to result in thicker bainite needles.
The parameter that most of all seems to control the thickness is the transformation temperature,
lower temperatures leads to thinner bainitic needles. [18]
The primary reason for the high strength of the martensite-bainite duplex structure is the
morphology of the lower bainitic needles which form during isothermal heat treatment at low
temperatures close to the Ms. [1] Enhanced mechanical properties due to formation of lower
bainite has only been observed, when lower bainite is formed by isothermal heat treating. Lower
bainite formed during continuous cooling only have impairing effect on the properties, when
combined with martensite. [26]
17
There is a variation ways to create bainite with different quenching mediums. [27][28] However,
in order to form isothermally transformed bainite, isothermal heat treatment in molten salt is a
well-established method. [2]
It can be difficult to distinguish the martensite and the bainte structures from each other. When it
comes to distinguishing tempered martensite and lower bainte, G.R speech et.al says in their
report on tempered steel that martensite will get a body central tetragonal lattice arrangement
when the carbon level is more than 0.2 %. Under the 0.2% carbon limit it will have a BCClattice. [46] The BCT-lattice is a consequence of trapped carbon atoms. When trapped C atoms
are allowed to diffuse and precipitate and form carbides, it could possibly lead to that the
martensite BCT lattice will transform and have more uniform lattice parameters similar to BCC.
The lower bainite structure consists of ferrite (BCC) with fine dispersed carbides precipitated
inside and between the ferrite plates. This also applies to the tempered martensite structure.
However, one difference of the two structures is how the carbides are orientated inside the ferrite
or tempered martensite phase. The intralath precipitated carbides in lower bainite have an
orientation of 55-65 deg to the long axis of the bainite plates. The tempered martensite has
carbides with multiple habit orientations, while lower bainte have carbides with a single habit
orientation. This distinction can only be seen when studying the microstructure in a SEM or
TEM [47].
2.5.
Banded structures
Banding has been a problem in AISI 4140 steel and can have negative effects on the mechanical
properties resulting in microscracks between areas with different harnesses.
Banding is a phenomenon in steel materials where fiber lines can be seen in the microstructure.
Dark and light colored fiber lines are formed during etching due to variation in austenite
transformation products in different bands.
The main cause of banding is inter-dendritic segregation during casting, and results in bands with
different chemical compositions, which consequently results in bands with different austenite
transformation products. Even though the chemical composition varies, it does not necessary
lead to banding. If the cooling rate is rapid enough the whole sample will transform into
martensite, which will provide a homogenous microstructure. But, the segregation can result in
different martensite band with a variation in etching color due to the variation in chemical
composition.
Different elements are more prone to segregation. Solute elements with high redistribution
factor/partition ratio (k) will have higher tendency to segregate. One other important factor,
18
which will influence the redistribution, is the amount of the segregated element. For example,
phosphor (P) has a high k value but is often present in low amounts, because it is removed by a
refining step during secondary steel making, due to the impairing effects it has on the mechanical
properties. Manganese (Mn) on the other hand has a relatively low k, but is often added in higher
concentrations, and will therefore have a bigger impact on the segregation. Certain elements will
also influence other elements redistribution. Mn together with chromium (Cr) is strong carbide
formers, and will affect the redistribution of C, by lowering the activity of C. Regions with high
Mn and Cr content will therefore attract C, in order to increase the local C concentration.
Elements such as nickel (Ni), Silicon (Si) and P will have an opposite effect on the C activity,
and will reject C from areas with high concentrations of these elements. [6] The affinity between
C and Mn, Cr, will make it very difficult to homogenize the material via normalization after
forging. Even if the high temperature will lead to high carbon diffusion the diffusion of Cr and
Mn is slow, and will attract the C, resulting in very little or no homogenization. Homogenization
at a high temperature during a long times is necessary to reduce the inhomogeneous composition,
see Figure 11, which shows how the distribution of Mn depending on homogenization
temperature and time. The same relationship can be assumed to be relevant for Cr and other
strong carbide formers. [41]
19
Figure 11. . The variation in strong carbon binder concentration after homogenization depends on the homogenization
time and temperature. In this picture the Mn concentration is plotted against time and temperature. [42]
In the work of Garcia Navas et.al the authors claim that more fiber lines indicate a higher
material flux which is a result of higher plastic strain. [41] The banded structures indicate
inhomogeneity, and difference in deformation of the bar stock, and regions with higher density
of bands is a result of higher more severe deformation.
20
2.6.
Heat treating
The process of creating a martensitic structure is done by rapid quenching, fast enough to avoid
the C-curves in the TTT diagram, see Figure 12. The C-curves represent diffusional
transformation of austenite into pearlite and bainite. The area between y-axis, representing
temperature, and the tip of the C curve is the shortest amount of time needed for austenite to
transform into either pearlite or bainite, and is called the incubation time. Transformation will be
fastest when the diffusion and the driving force for formation are both high. High incubation
time will aid transformation to martensite, and can be increased by higher alloy content. Forming
martensite through a fast quench is often performed on steels with more than 0.3 % C due to the
fact that the increase in hardness is more significant for steel grades with carbon content equal to
or higher than 0.3wt%. [6]
Figure 12.Is a picture of isothermal heat treating diagram of AISI 4140. The C-curves represent at which temperatures
the diffusion controlled transformation of austenite is most rapid. Arrows point out the C-curves. [29]
At Cummins Fuel Systems, the products used in the fuel injection systems are quenched in order
to form a martensitic structure. This has been a way to ensure certain properties including high
hardness and toughness. The higher carbon content in austenite before transformation into
martensite has an increasing effect on the hardness. There is an increase in dislocation density
and a higher distortion of the BCT lattice, due to more trapped carbon atoms. The high hardness
will consequently result in lower toughness and strength. Low toughness is mediated by a
process called tempering, where the carbon atoms are allowed to form carbides and the
21
dislocation density is reduced. [23] However, research has shown that bainitic structures can
have as good as or even better properties than tempered martensite. [22] Research reports that
bainite is formed through quenching with different mediums such as gas, water or isothermal
treatment in molten salt. [30][31] A variation of quenching speeds and mediums will be tested to
better understand the formation of a mixed structure of bainite and martensite.
In order to design a heat treating process that produces a mixed microstructure, there are a
variety of parameters that will have an effect, and needs to be taken in to consideration. The
alloy content is one important factor, increasing the time austenite can be under the eutectic
temperature without transform into other structures, resulting in a slower transformation. The
foreign alloy atoms are occupying iron lattice sites, highly influencing the carbon mobility.
Moreover, lower carbon mobility provides a better hardness profile, with a more even distributed
hardness. [23] In order to control the final microstructure it is necessary to control the cooling
rate. If the cooling rate is slow, the risk that austenite will be transformed into ferrite and pearlite
is high. Faster cooling rates will suppress the diffusion of carbon, allowing austenite to transform
into bainite and/or martensite. [23]
2.6.1. Quenching
During the quenching process the transportation of heat from the steel component is the deciding
factor. The rate of the transformation of heat from the steel part will decide the amount and type
of transformation products and consequently final microstructure. The heat transport can be
divided into two different phenomena: conduction and convection. Conduction is the transport of
heat from the interior of the sample to the surface and convection is the transport of heat away
from the surface by the quenching media.
The quenching process can be divided into three different steps:
A layer of vapor/steam is created close to the steel surface. The steam layer insulates the
surface resulting in a low initial cooling rate.
The vapor layer is broken down and the fluid will come into contact with the surface.
The water then evaporates and forms bubbles, which are transported away from the
surface. The cooling rate is rapid during this stage.
When the surface temperature drops to a temperature below the boiling point of the fluid,
the cooling rate is determined by the transportation of heat through either conduction or
convection, resulting in a low cooling rate. [30]
The different steps are illustrated in Figure 13.
22
Parameters that influence quenching are: The circulation of the quenching medium, carbon and
alloy content, austenite grain size and component thickness. [30]The agitation level will highly
affect the three quenching steps and how long they are present. [23]
Research has shown that interrupted quenching can greatly impact the final product when it
comes to phase transformation, distortion and residual stresses. Interrupted quenching makes it
possible to manipulate the diffusion of carbon. Which enables the opportunity to control the
cooling curve, and thereby the amount of different austenite transformation products. Isothermal
heat treatment allows carbon diffusion to be somewhat controlled, thereby lowering the stresses
in the material and resulting in fewer process steps such as tempering.
Figure 13. Different quenching stages in liquid quenching. A vapor blanket will be formed around the quenched part
initially. When the temperature decreases bubbles will be formed on the surface and transported away from the surface.
When the temperature has deacreses further the transport of heat from the part will be controlled by convecttion. [30]
2.6.2. Oil quenching
Oil is commonly used as a quenching medium in order to get even and consistent mechanical and
metallurgical properties. Oil quenching controls the heat transfer during rapid cooling and
increases the wetting of the steel surface, limiting any thermal and transformational gradients
which can cause distortion and cracking. Oil quenching is also beneficial because it is possible to
predict the distortion pattern due to somewhat consistent heat transfer with similar geometries
[30]
2.6.3. Gas quenching
During gas quenching, inert gas is blown over the steel surface. Heat is absorbed by the gas and
is transported to water cooling heat exchangers, and then recycled back to the system for reuse.
The process can be varied by controlling the temperature, velocity and pressure of the gas. Also,
23
the type of gas that is used will have a large impact on the transportation of heat form the
quenched part. Inert gases are commonly used because no chemical reactions with the surface
take place, and therefore post-heat treatment surface finishing is not necessary. Typical gases
used in quenching are Argon, Helium, Nitrogen and Hydrogen. [23]
Gas quenching is able to achieve higher cooling rates compared to cooling in still air. Even
though quenching with gas is slower than with oil, gas is used to reduce the risk of high
distortion and cracking. Gas provides a more uniform hardness and is cleaner compared to liquid
quench mediums. [23] Gas is often used for components with large sections, when a more
uniform hardness profile is necessary. Components with thinner cross sections can be oil
quenched because the boiling stage decreases with thickness, and the temperature where boiling
takes place increases as the geometry decreases, and can provide a relatively even hardness
2.6.4. Molten salt quenching
Quenching in molten salt is a popular way of attaining martensite, it allows for a more controlled
cooling process, compared to conventional quenching methods e.g. in water, oil and gas.
Quenching in molten salt is particularly applicable for steels with high hardenability. The risks of
distortion, non-uniform hardness, thermal stresses and transformation stresses are decreased
when quenching in molten salt.
Austempering is a cooling process where the steel is cooled rapidly in molten salt in order to
prevent pearlite transformation. It allows isothermal heating at a low enough temperature to
allow decomposition of austenite into bainite, due to the fact that molten salt can sustain higher
temperatures without degrading compared to most quench oils [31] [32]. Quenching and
isothermal heat treatment in molten salt enables isothermal heat treatment between the pearlite-C
curve and the MS-temperature. Another benefit of quenching in molten salt is that the vapor layer
does not form and therefore results in more rapid cooling at higher temperatures. [32]
Three different features in the TTT-diagram are necessary in order for austempering or
martempering to be feasible:
A low enough MS temperature to allow bainite formation
High hardenability so the formation of pearlite can be prevented
Reasonable transformation times to bainitc structure, not too long. [33]
Due to its negative impact on mechanical properties, it is important to ensure that no upper
bainite is formed; therefore it is necessary to isothermally heat treat at a temperature close to the
MS-temperature. [31][32]
24
2.6.5. Isothermal heat treating temperature
This project is on isothermal heat treating closely located around the MS-temperature. The idea is
to try to form lower bainite, by performing isothermal heat treatment as close to the Mstemperature as possible without going below it. This idea shaped the isothermal heat treating in
the first trial, Trial A. However, there have been rapports on lower bainite formation under MS.
The reason is said to be, that thermal stabilization of the austenite phase take place initially when
the temperature reaches just below MS, at a so called arrest temperature Ta. This phenomenon
will interrupt the martensite formation. The thermal stabilization will allow lower bainite
formation, and will appear as thin needles in the material. Earlier work have indicated that this
could also lead to thinner lower bainite needles than formed above MS, and cloud possible result
in better refining of the martensite packets. [43] Also, the formation of lower bainite seems to be
slower due to the lower transformation temperature, and could therefore be better controlled. [1]
In steels with low carbon content the Ms-temperature can be relatively high resulting in the
possibility of redistribution of carbon atoms during quenching, before reaching room
temperature. This phenomenon is referred to as autotempering or quenchtempering. Carbon
atoms segregate further during quenching to high energy areas e.g. areas with fine dislocation
structure or to packet/lath boundaries, resulting in precipitation of cementite particles. Moreover,
low cooling rate under Ms could result in so called autotemering. If the Ms-Temperature is high
autotempering is consequently more difficult to prevent. [6] Autotempering is purposely used in
some processes to temper the martensite, and similar to austempering reduce the amount of
distortions and stresses. This reduces the necessary production steps, because tempering after
quenching is not needed.
One way to prevent autotempering can be to use a lower isothermal heat treating temperature.
2.7.
AISI 4140
The AISI 4140, also referred to as Chrome-moly steel, have good strength to weight ratio. 4140
is considered strong and have relatively high hardenability. Fully hardened AISI 4140 steel can
reach 60 HRC. High hardness is related to the carbon content; higher carbon content will result
in higher hardness and consequently higher wear resistance. AISI 4140 is used in systems for
pressurized tubes, such as fuel injection bodies. Such systems create high pressure, which pulses
up and down over long periods of time. The 4140 is easy to case harden in order to protect
against wear and erosion. Creating a thin case by nitriding can increase the fatigue strength with
up to 30%. The material used in this work is provided by Ovako AB. The composition
restrictions are presented in Table 1. [36][37]
25
AISI 4140 can experience different kinds of embrittlement during or after heat treating e.g. Blue,
temper and martensitic embrittlement.
Blue embrittlement is associated with precipitation hardening. The steel will experience increase
in strength and a decrease in ductility, resulting in a brittle material. The name blue
embrittlement, is related to that it occurs between 230-370 C which is called the blue heat range.
Tempering embrittlement is caused by precipitates of trace elements such as Mn and Cr. Local
depletion of alloy elements at prior austenite grains resulting in intergranular cracks, along the
prior austenite grain boundaries.
Martensite embrittlement is a result of two main phenomena; segregation of impurities e.g. P, to
the grain boundaries during austenizing and precipitation of cementite particles between the
martensitic laths during tempering. [38]
The composition range of different elements for AISI 4140 from Ovako AB can be seen in Table
1.
Table 1. General compositon range of AISI 4140 (wt%) at OVAKO AB
Element
AISI 4140H
(326)
2.8.
C
Si
Mn
min
0.38
0.2
0.8
max
0.43
0.35
1
P
0.015
S
Cr
Ni
Mn
0.005
0.8
0
0.15
0.01
1
0.2
0.25
Etching
In order to study a materials microstructure it is necessary to etch the mounted samples. Different
etch-methods and etch-chemicals will help the researcher to reveal different information about
the configuration of the microstructure. To be able characterize the microstructure is an
important work to understand materials mechanical properties, which is related to the materials
grains size, defects and the variation of formed transformation products.
The most common way of etching is to use a so called attack etch method. Attack etching is
basically controlled corrosion where the different features of the microstructure will etch faster
than other parts. The microstructure together with the etching chemical can be seen as an
electrochemical cell. Different structures and phases have different electrochemical potential.
Features with higher electrochemical potential will easier be etched.
Pearlite for example consists of ferrite and cementite. Ferrite has a higher electrochemical
potential than cementite. The ferrite will act as an anode and the cementite structure as a cathode,
resulting in a deeper etch of the ferrite phase. The deeper etch of the ferrite phase will result in a
darker appearance, due to the difference in the reflection of the incoming light, when the surface
26
is studied under a light optical microscope. The bainite structure also consists of ferrite and
cementite, but with a non-lamellar and a finer distribution of the two structures. There are a
number of ways to reveal the bainite structure but the one most recommended is picral because it
attacks the ferrite carbide interface. Nital do not etch two phase structures as good as picral.
Nital attacks the ferrite grain boundaries and is used to reveal ferrite grains and as-quenched, untempered martensite. [39]
One alternative way of etching is by using a tint etching chemical. A tint etchant will instead of
removing material as the attack etch do, deposit a film on top of the surface. The most common
tint etchants are sodium metabisulfite or potassium metabisulfite. In this study the sodium
metabisulfite is going to be used. The sulfide ions in the etchant will act as reactants and will
form a film on the surface. Tint etching is a good way of revealing martensite in a mixture of
structures. [39]
A mixture of attack and tint etchant is a way to better reveal the microstructure. The attack
etchant will remove the oxide layer that forms on the surface when exposed to air. It is necessary
to remove the oxide film in order for the tint etchant to etch uniformly over the whole surface.
The sodium metabisulfite will then deposit a layer of sulfide on the surface. [39]
2.9.
Point Counting Method
In order to count the amount of different structures in the material, pictures of the microstructure
is going to be taken at 500x magnification. A point counting method is adopted from the two
dimensional systematic point count method described more in detail by Hilliard and Cahn. [40]
A pre-decided grid is used and is superimposed on the picture of the microstructure. Each corner
in the grid/lattice represents a point. The amount of points which ends up inside the analyzed
features, divided by the total amount of points, provides number of the percentage of the
analyzed phase/structure. A schematic picture can be seen in Figure 14.
27
b)
a)
Figure 14. Scematic picture explaining the point counting meathod used in this study. The number of points whitin the
features analysed divided by the total amount of point provide a procentage.
The grid of points should be coarse enough to ensure that not two adjacent point end up in the
same feature. One other way of defining the necessary point space, is that the minimum distance
between two adjacent points most be greater than any caliper radius of the studied features. [40]
3. Materials
The material is provided by Ovako AB and is AISI 4140 steels. Two materials, 326 F and 326 C,
are going to be tested. The 326C is the standard 4140 steel from Ovako AB, while the 326 F is
designed to have a higher hardenability. The difference in hardenability is due to the increased
levels of Cr and Mo in 326F steel. Chemical composition for material 326F and 326C can be
seen in Table 2 and Table 3 respectively.
Table 2. Chemical composition (wt%) Ovako steel grade 326F
C
Si
Mn
P
S
Cr
Ni
Mo
Cu
V
Al
0,42
0,25
0,94
0,011
0,006
1,05
0,12
0,23
0,126
0,008
0,027
Table 3. Chemical composition (wt%) Ovako steel grade 326C
C
Si
Mn
P
S
Cr
Ni
Mo
Cu
V
Al
0,42
0,24
0,95
0,009
0,009
0,99
0,14
0,17
0,15
0,009
0,027
The 326 F and C arrived as peeled bars and had been spheroidized annealed and hardened
tempered respectively. Pictures of the microstructures can be seen in Figure 15.
28
b)
a)
Figure 15. The microstructures of both materials were studied before heat treating. The two materials were heat treated
differently at the supplier. a) Quenched and tempered 326 C x500, b) spheriodized annealed 326F x500. Both were
etched with 2% nital.
The structure is finer in the 326C quenched and tempered material, while the 326F spheriodized
annealed material has a coarser microstructure.
29
4. Experimental procedures
4.1.
Sample preparation
The bar stock was cut into pieces of approximately 10 and 20 mm, see Figure 16. Three samples
for each material, heat treatment and sample thicknesses were prepared. 326C and 326F have
diameters of 36.5mm and 41mm respectively.
a)
36.5 mm
b)
Figure 16. The bar stock were cut in samples with two different thicknesses, 10mm and 20mm. The picture shows the two
different thicknesses of materila 326C with a diameter of
One other samples thickness is added to the heat treatment using the vacuum heat and gas
quench method, the thickness is 45mm.
4.2.
Sample analysis
In order to study the microstructural changes and hardness after heat treatment the samples
where sectioned to investigate in the longitudinal direction, see Figure 17.
Figure 17. A Schematic picture showing the surfaces studied of each sample. The samples were investigated in the
longitudinal direction of the bar stock.
30
Both micro and macro hardness tests were performed. Macro hardness test was taken on the
surface of the samples, while the micro-hardness test was taken on the mounted surface of the
samples. The microhardness readings were taken over the whole surface to see the hardness
variation from the surface to the core of the sample. Five readings were taken on multiple
locations on the surface, at the core and between the core and the surface, in order to study the
homogeneity of the microstructure. An illustrative picture can be seen in Figure 18.
Core
Surface
Figure 18. Schematic picture showing were the microhardness reading were taken. To better understand the variation in
hardness between the surface and the core, readings were taken over the whole mounted surface.
The result was studied using well known statistical tools such as 2 sample T tests and one way
anova tests, in order see any statistical differences between Trails, materials, sample thickness,
radial hardness variation and time isothermally heat treated.
31
4.3.
Heat treatment
Three heat treatments are going to be tested to study the bainite transformation in AISI 4140
steel. The heat treatments are further described in Table 4.
Table 4. Heat treatment performed in this work.
Designation
Vacuum heating and as quenching
Heat Treatment
Austenitizing at 843ºC at 1.5 h in an
atmosphere 100 micron vacuum, gas
quenching (Nitrogen) at 2.758 barr down to
room temperature.
Austempering, Trial A: ( Isothermal heat
1. Preparation step*
treating T=343ºC)
2. Austenitizing in molten salt at 871ºC
for 1h and placed in quench salt at
343ºC for 60-120 s, then immersed in
water for 1 min.
Austempering, Trial B: ( Isothermal heat
1. Preparation step *
treating T=320ºC)
2. Austenitizing in molten salt at 871ºC F
for 1h and placed in quench salt at
320ºC for 60-120 s, then immersed in
water for 5 min.
3. Tempered at 200ºC for 2h
* parts evenly spread out in basket and dried in 260ºC
for 15 min
A schematic picture of the heat treatment cycle for Trial A and B can be seen in Figure 19. The
difference between Trial A and B is the isothermally heat treating temperature, 343ºC and 320ºC
respectively, and that a subsequent heat treating step is added to Trial B. The tempering step in
Trial B is adopted to limit the risk for tempering embrittlement.
32
Figure 19. Shcematic picture of the heating cycles. Trail A have a isothermal heat treating temperature higher than Trail
B, and Trial B have a temepering step. The fisrt step represents when the samples were dried before heat treating.
4.4.
Etching methods
A variation of etching methods is going to be tested in order to study the variation of the
appearance of the different structures when studied under a LOM. This is done in order to test
and map the different etching opportunities and to do preparatory work in order to determine an
etching process to better reveal, and distinguish as-quenched and tempered martensite from
lower bainite. The chemicals used in this study can be seen in Table 5.
Table 5. Tested etching chemicals.
Etchant
2% Nital
Description
Attack etchant. One step process.
4%Picral + 5% Sodium metabisulfite in water
Attack etchant + tint etching. Mixed; the
etching process consists of one step.
4%Picral+HCl
Attack etchant + surface agent. Mixed, the
etching process consists of one step.
4% Nital + 5% Sodium metabisulfite in water
Attack etchant + tint etchant. Not mixed, two
step etching process.
33
5. Result and Discussion
5.1.
Vacuum heating gas quenching (VHGQ)
5.1.1. VHGQ test 1
Three samples with three different thicknesses were heat treated at 1550 F in vacuum (100
microns), and quenched in gas at 40 psi, in an atmosphere of nitrogen.
The banding is quite large, which indicates poorly or no homogenization after casting. Banding
is caused by segregation, resulting in areas with difference in alloy concentrations, as seen in
Figure 20.
Figure 20. Banded strucutre etched in 2% Nital.
In bands, with light colored structures, it is possible to see manganese-sulfur (MnS) particles,
this proves that higher amount of Mn is present in these areas, as seen in Figure 21. Mns particles
cannot be spotted in the darker bands. High amount of Mn would consequently lead to higher
amount of C. The high Mn and C concentrations in these areas would therefore motivate the light
colored structures as being martensite and lower bainite. Increasing Mn and C content pushes the
C-curve in the TTT-diagram to the right, increasing the incubation time. Higher incubation time
will support the martensite and lower bainite transformation. If the cooling is rapid enough and
the sample is transformed into martensite, no band will occur, except for bands caused by the
etching due to the different solute content the different martensite regions.
34
a)
b)
c)
d)
Figure 21. Pictures taken in LOM. a) Core with low magnification, b) core with high magnification, c)
surface with low magnification and d) surface with high magnification
The macro hardness test revealed a higher hardness at the perimeter of the sample. The average
macro-hardness is plotted against the samples thickness and can be seen in Figure 22. The
decrease in hardness with increasing sample thickness indicates more non-martensitic
transformation products in thicker samples.
Average hardness vs. sample thickness
60
55
HRC
50
45
Macro-hardness
40
Micro-hardness
35
30
0
10
20
30
40
50
Thickness (mm)
Figure 22. Avereage macrohardness and mirohardness ploted aginst sample thickness.
35
All three samples were quenched the same way in the same heat treating run. The variation in
hardness between the three samples is almost linear. Samples with 10 and 20 mm thickness were
both located around the 50-55 HRC specification range. But the variation between the highest
and lowest value is >5 HRC, which is too much. The variation should not be higher than 5HRC
because it could result in microcracks in the material. An example of such a region is the
boundary between two bands. This is due to the variation in hardness between bands, resulting
from the variation in chemical composition and consequently different hardenability.
It is possible to see some lower bainte structure in the images. But it is also possible to see ferrite
and pearlite structures. Higher gas pressure initially during quenching can possibly reduce the
amount of formed ferrite and perlite, as-well as increase the agitation of the gas initially and then
reduce the pressure and shut down the fan responsible for the agitation after a couple of seconds.
This could possibly provide better result when it comes to reducing the ferrite and pearlite
formation and support the lower bainite formation.
The formation of thin lower bainite needles seems to be necessary in order to result in a
refinement of the martensite structure and thereby an increase in strength. The formation of
needles is dependent on the temperature, low temperature is necessary. Moreover the needles
need time to grow in order to partions the austenite grains. If this is possible by conventional
gas quenching is not clear.
It should also be mentioned that an enhancement of the mechanical properties caused by lower
bainite needles have not been seen in conventional quenching. Isothermal heat treating is
necessary in order for this phenomenon to take place.
5.2.
Austempering in molten salt
5.2.1. Trial A
Similar banded structure as seen in the vacuum heated and gas quenched samples can be seen in
the samples isothermally heat treated in molten salt. The banding seems to be most severe in the
center of the pieces, which indicates more segregation, and thereby a higher plastic strain in the
center of the samples. In the work of Garcia Navas et.al the authors claim that more fiber lines
indicate a higher material flux which is a result of higher plastic strain. [41] The banded
structures indicate inhomogeneity, and difference in deformation of the bar stock, this is further
discussed in section 2.5.
The purpose of the austempering in molten salt was to try to control the levels of formed bainte
in a matrix of martensite. The banded structure makes it difficult to control the levels of bainite
36
and martensitic structures. The amount of formed ferrite and pearlite, if any, is also very difficult
to predict, when the composition of the material is different depending on banded zones. High
initial quenching rate is necessary in order to ensure that no pearlite and ferrite is formed.
Also, due to the segregation during solidification and the resulting banded structure. The
variation of alloying elements in these bands could also affect the MS-temperature for each band
resulting in different amounts of bainite and martensite transformation products. This seems to
be the case for most samples in Trial A. One way to test this could be to investigate the amount
of retained austenite in each band.
The thinner samples showed a slightly higher surface hardness then the thicker samples, all
materials and both thicknesses showed a peak in hardness after 90 s. Curves for each sample
type are compared to each other in Figure 23. Material 326F has a higher surface hardness than
material 326C. For the same thicknesses but different material, the average difference is around
1,4HRC. This probably a result caused by the higher Cr and Mo content in the 326F material
resulting in a higher hardenability and thereby a higher surface hardness.
HRC
Macro Hardness
45
44
43
42
41
40
39
326C 10mm
326C 20mm
326F 10mm
50
70
90
110
130
326F 20mm
Time (s)
Figure 23. Hardness plotted against time in isothermal heat treatment. Different lines represent different material and
thickness combinations.
Macrohardness result for each sample and time isothermally heat treated together with the
standard deviation for Trial A can be seen in Table 6. The data is collected from all 36 samples,
three readings from each sample.
37
Table 6. Macrohardness results from Trial A.
Sample
326C 10mm
326C 20mm
326F 10mm
326F 20mm
Macrohardness (HRC)
60s
90s
120s
40.87 ± 1.73 43.74 ± 0.63 43.52 ± 0.60
39.57 ± 0.62
40.5 ± 1.40
39.96 ± 2.38
42.49 ± 1.17 44.52 ± 1.02
44.19 ± 0.57
40.66 ± 0.29
43.32 ± 0.60 41.51 ± 1.15
The hardness restriction for this study is set to be within 50-55 HRC. A larger variation in
hardness can lead to microcracks. A boxplot of each sample variation can been in seen in Figure
24. Each box is an individual combination between: material (326C or 326F), samples thickness
(10mm or 20mm) and time isothermally heat treated, (60, 90 or 120s).The dotted lines divide the
graph into different time zones. The red reference lines show the 50-55 HRC restriction. Each
box plot includes 15 data points.
Boxplot of Microhardness (HRC), Trial A
65
Microhardness (HRC)
60
55
55
50
50
45
40
Thickness (mm):
Material:
Time (s):
10 20
326 C
10 20
326 F
10 20
326 C
60
10 20
326 F
90
10 20
326 C
10 20
326 F
120
Figure 24. Boxplots of microhardness in Trial A.
The variation between the highest and lowest value is large and can be a result of the banded
microstructure together with the expected surface to core variation, due to the relatively low
38
hardenability. The microhardness shows similar pattern as for macrohardness data except for the
326 F 10 mm, which have a lower hardness in relation to the other sample variations. More data
have to be collected to further understand why sample 326F 10mm deviate. Average
microhardness for each sample variation are plotted against isothermal heat treating time, see
Figure 25.
Micro Hardness
HRC
49
47
326 C 10mm
45
326 C 20mm
43
326 F 10mm
41
326 F 20mm
50 60 70 80 90 100 110 120 130
Time (s)
Figure 25. Mirohardness plot for eacth sample variation and time isothermally heat treated at 650F.
Microhardness result for each sample and time isothermally heat treated, together with the
standard deviation for Trial A can be seen in Table 7.
Table 7. Microhardness results from Trial A.
Sample
326C 10mm
326C 20mm
326F 10mm
326F 20mm
Microhardness (HRC)
60s
90s
120s
44.69 ± 1.81 49.62 ± 3.80
49.2 ± 4.69
42.46 ± 2.30
47.23 ± 2.48 42.79 ± 3.14
47.17 ± 2.91
48.03 ± 3.34 45.45 ± 3.25
43.65 ± 2.59
49.01 ± 4.92
44.52 ± 3.08
The 326 C seems to have some ferrite and pearlite in its microstructure after quenching. Ferrite
and pearlite is not present in the same extent in the 326F material after heat treatment, see Figure
26. These pictures are representative for each material and thickness combination. Moreover,
less non-martensitic transformation product is present in the 326F material, which is an
indication that the higher hardenability, increase the incubation time, which makes it easier to
create less amount of lower bainite structure. Moreover, less lower bainite content result in more
39
as-quenched martensite and consequently higher strength. All this is to be expected, due to the
difference in alloying composition.
a)
b)
c)
d)
Figure 26. Representive images taken of the microstructres at x500 magnification after 60 s. a) 326 C 10mm b) 326 C
20mm, c) 326 F 10mm and d) 326 F 20mm
The microstructure before heat treating may have provided different circumstances resulting in
variation in banding and amount of formed bainite. Material 326F had a more course
microstructure before heat treating maybe it contributed to the less formed lower bainite in
material 326F. The microstructure is finer for 326C than for 326F.
5.2.2. Isothermal heat treating temperature
The smallest amount of transformed bainite in Trial A was above 50%, lower isothermal heat
treating temperature could possibly lead to less transformed lower bainite. Less lower bainte
formation could provide better understanding of the refining effect that the lower bainite needles
will have on the mechanical properties. The goal should be to reach the 25-30% lower bainite
content that in earlier work have shown to enhance the strength of high-tensile steel. One way to
reach this goal could be to lower the isothermal heat treating temperature.
Because of the presence of tempered martensite in the structure, it is likely that the cooling under
Ms has not been fast allowing carbon to diffuse. Because no subsequent tempering process was
performed in the Trial A the tempering of the produced martensite had to be tempered during the
40
quenching process. The isothermal heat treatment was conducted above M S, which indicate that
the cooling from isothermal heat treatment T down to room T was relatively slow.
It was decided to use an isothermal heat treating process under MS and a subsequent tempering
step, due to a variation of factors:
It is difficult to distinguish the tempered martensite and the bainite structure from each
other. In order to test etching methods to distinguish the tempered martensite and the
bainte structure, it was decided to temper the samples in order to temper all martensite,
and to see the difference between as-quenched and tempered martensite. The literature
provides various methods to do reaveal as-quenched marteniste. But in order to better
understand what separates the appearance of lower bainite and tempered martensite a
tempering step was included in the heat treating plan for Trial B.
One other reason to add a tempering step is to level out the hardness profile in the
samples. It was shown in Trial A that the variation within the samples from the core to
the surface was quite large. In order to ensure that no brittle as-quenched martensite was
still present after heat treating, a subsequent heat treating step was introduced.
In order to slow down the growth of the bainite structure, the isothermal heat treating
was decided, in Trial B, to be performed under MS. Hopfully resulting in less amounts of
formed lower bainite.
Lower isothermal heat treating temperature could possibly lead to an increase in
quenching speed.
The tempering process is going to be performed on the boundary for temper embrittlement. AISI
4140 experience temper-embrittlement if tempered between 393 and 700 F.
Bainite formation under MS has been studied before and the heat treating process was influenced
by heat treating Trials designed by Tomita and Okabayashi [1]. The samples were isothermally
heat treated at the same times used in Trial A, between 60 and 120 seconds, also adopted from
the work from Tomita and Okabayashi. [1]
5.2.3. Trial B
A boxplot of the microhadrness in Trial B can bee seen in Figure 27. The microhardness is more
evenly distributed between the core and the surface of the samples in Trial B compared to Trial
A. Which indicates that the tempering process after isotheral heat treating had a possitive effect
on the hardness variation. The pattern seen in Trial A where the 326F material had a overall
higher hardness than 326C is lost in Trial B, and can be a result of the subsequent tempering
step. The average microhardness of each sample variation can be seen in Figure 28. When the
samples were etched and studied under a microscope material 326F provided a better
41
microstrucutre when it comes to less or no pearlite and ferrite transformation products, which
further indicates that the lower bainite and the tempered martensite have a simmilar hardness.
Boxplot of Microhardness (HRC), Trial B
Microhardness (HRC)
60
55
55
50
50
45
40
Thickness (mm):
Material:
Time (s):
10 20
326 C
10 20
326 F
10 20
326 C
60
10 20
326 F
10 20
326 C
90
10 20
326 F
120
Figure 27. Box plot of the Microharndess for Trial B.
HRC
Microhardness
51
50
49
48
47
46
45
44
43
326C 10mm
326C 20mm
326F 10 mm
326F 20mm
50
70
90
110
130
Time (s)
Figure 28. Mirohardness plot for eacth sample variation and time isothermally heat treated at 608F.
42
Microhardness result for each sample and time isothermally heat treated, together with the
standard deviation for Trial B can be seen in Table 8.
Table 8. Microhardness results from Trial B
Microhardness (HRC)
Sample
60s
90s
120s
326C 10mm
47.18 ± 1.26
47.19 ± 2
48.57 ± 1.08
326C 20mm
43.56 ± 2.04
47.06 ± 4.81
46.72 ± 3.55
326F 10mm
326F 20mm
50.51 ± 1.50
50.66 ± 2.42
44.70 ± 1.86
44.45 ± 1.31
46.70 ± 1.21
48.14 ± 2.19
Representative pictures for each material and thickness combination can be seen in Figure 29.
The amount of non-martensitic transformation products is lower in the 326F than in the 326C
material. It is reasonable to relate less amount lower bainite formed in material 326F to the
higher hardenability. The typical needle like appearance of the lower bainite can be seen in
Figure 29 c) and d).
a)
b)
c)
d)
Figure 29. Representive images taken of the microstructres at x500 magnification after 60 s, a) 326 C 10mm b) 326 C
20mm, c) 326 F 10mm and d) 326 F 20mm. Isothermally heat treated for 60 s in 608F
43
The microstructure for material 326C is finer than for material 326F. This makes it easier to
distinguish the structures form each other in material 326F, and can also be an influencing factor
affecting the amount of formed lower bainte.
5.2.4. Comparison of data
2 sample T tests were done in order to see if there was any statistical difference between
variations of parameters.
The first tests were done for Trial A, the test was done in order to study the martensite
transformation because it was difficult to count and classify any other structural features in
material 326C.
There was a significant difference between the two materials and a box plot of the %martensite
for each material can be seen in Figure 30. The 95% CI is between 6.3-24% differences in
martensite content between the two materials.
Boxplot of Martensite%
60
Martensite%
50
40
30
20
10
0
326C
326F
material
Figure 30. Box plot showing the difference in %martensite between material 326C and 326F
Same significant difference can be seen when studying the difference in macrohardness between
the two materials, see Figure 31. Material 326F have approximately 1.4 HRC higher
macrohardness than material 326C. The difference in macorhardness and in %martensite
44
supports each other, because it is reasonable to assume a higher hardness at higher martensite
levels.
Boxplot of Macro hardness HRC (surface)
46
Macro hardness HRC (surface)
45
44
43
42
41
40
39
38
37
326C
326F
material
Figure 31. Box plot showing the difference in macrohardness between material 326C and 326F
Difference in microhardness for Trial A and B dependent on parameters such as: material, Time,
thickness and orientation can be seen in Table 9. Also, difference in microhardness between the
two trials can be seen in Table 9.
45
Table 9. 2 sample T and one way anova tests to compare different parameters with each other see if there was any
statistical difference in microhardness.
Thickness (10
and 20mm)
Radial variation, (surface,
middle and core)
Mean
Microhar
dness
(HRC)
10mm> 20mm,
2,4 HRC
average
difference
Surface & middle > Core
46.2 ± 4
10mm> 20mm,
1,2 HRC
average
difference
Surface & middle > Core
47.1 ±
3.1
Difference in Microhardness (HRC)
Material(326C
and 326F)
Tria
lA
No significant
difference
Tria
lB
No significant
difference
Time (s), (60,90
and 120s)
90 s showed
highest values,
no significance
difference
between 60 and
120s
90 s showed
lowest values,
no significance
difference
between 60 and
120s
There was a difference in microhardness between the both thicknesses in Trial A. Samples with
10mm thickness had in average 2,4HRC higher hardness than the 20mm thick samples. When
comparing the different time isothermally heat treated, samples heat treated for 90 s had a higher
hardness than the samples heat treated for 60 and 120 s. However, no significant difference could
be seen between the two materials in Trial A.
There is a difference in macrohardness (surface hardness) in Trial A. However, no significant
difference could be seen in microhardness from the surface to the core between the two
materials.
The thinner samples had in average a 1,2HRC higher microhardness than the thicker samples in
Trial B. When comparing the different time isothermally heat treated, samples heat treated for 90
s had a lower hardness than the samples heat treated for 60 and 120 s. However, no significant
difference could be seen between the two materials in Trial A.
Samples in Trial B had in average higher hardness than samples in Trial A. This can be a result
of the lower isothermal heat treating temperature in Trial B, together with the longer time
immersed in water in after isothermal heat treating step in Trial B. It is therefore reasonable to
think that lower temperature and faster quenching in water after isothermal heat treating result in
higher hardness. The hardness of the martensite structure seems to be the controlling factor when
it comes to the total hardness. While the lower bainite structure up to a 25-30% could lead to an
increase in strength.
46
Material 326 F provides the best result when it comes to distinguish the lower bainite and
martensite structures. The lower bainite content is less, due to the higher hardenability, and is
therefore further studied.
Both heat treating trials performed in molten salt has been analyzed with respect to the bainite
content. One assumption is that the darker etched areas are lower bainite, while the lighter
colored areas are martensite This is based on the morphology of the constituents in the darker
areas, due to the resemblance in appearance of the darker structure compared to earlier work. All
etching methods were adopted from earlier work on the subject, and is further discussed in 5.3.
The result is presented in two sample groups, samples with 10 and 20mm in thickness
respectively for material 326 F and can been seen in Figure 32 and Figure 33.
a)
c)
b)
Figure 32. Samples studied with same material (326F) and thickness (10mm) but form different heat treatments, a) 61%
lower bainite, austempering from Trial B at 60 s, b) 70% lower bainite, austempering from Trial A at 90 s and c) 85%
lower bainite, austempering from Trial A at 120 s.
a)
b)
c)
Figure 33. Samples studied with same material (326F) and thickness (20mm) but form different heat treatments a) 44%
lower bainte, austempering from Trial B at 60 s b) 54% lower bainite, austempering from Trial A at 90 s and c) 77%
lower bainite, austempering from Trial A at 120s.
The bainite content and the standard deviation can be seen in Table 10. It should be mentioned
that more samples should be investigated to better pin point the bainite content for each heat
treatment. There is no statistically significant difference between longer times isothermally heat
47
treated and increased lower bainte content. It is necessary to perform more tests and to collect
and analyze more data, to better understand the influence of transformation temperature and time
isothermally heat treated. Each value is an average of three samples and three different areas for
each sample.
Table 10. Sample thickness and %Bainite and the related heat treatment, for material 326F
Sample
326F 10mm
326F 10mm
326F 10mm
% Bainite
61 ± 6.93
70 ± 19.6
85 ± 9.23
Heat Treating
Trial B, 60 s
Trial A, 90 s
Trial A, 120 s
326F 20mm
326F 20mm
326F 20mm
44 ± 3.61
54 ± 10.38
77 ± 13.75
Trial B, 60 s
Trial A, 90 s
Trial A, 120 s
The lower bainite needles seem to be finer in Trial B, and can be a result due to the lower
transition temperature. Research has indicated that the lower bainte needles will not be provided
with an increased thickness at longer transformation times. The needle thickness depends more
on the transformation temperature, lower temperature provides thinner lower bainite needles.
[18] It would therefore be interesting to further study even lower transformation temperatures,
under the 320ºC temperature tested in this work, for lower bainite formation. How far below the
MS is it possible to isothermally heat treat and still form lower bainite needles? Lower
temperature could also help to increase hardness, due to increased quenching speed.
5.3.
Etching methods
Different etching methods where tested in order to try to find a method to better distinguish the
martensite and the lower bainite structure. Representative images of the different results can be
seen in Figure 34. All four methods reveal the martensite and non-martensite features. However,
some of the methods revealed more information. To choose etching method the results were
compared to earlier work on etching marteniste and bainite structures.
48
a)
b)
c)
d)
Figure 34. Variation of etching methods and chemicals: a) 2% Nital, b) 4% picral+ 1%Sodium Metabisulfite c)
4%Picral+HCl and d) 2%Nital and 5%Sodium Metabisulfite
Form the work of Vander Vort [44] upper bainite and as-quenched martensite have been etched
with sodium metabisulfite, and can be seen together with a picture of material 326F
isothermally heat treated for 120 s, in Figure 35. The as-quenched marteniste has been given a
brown color, while the bainite and other non-martensitic transformation products, if any, have a
dark blue color.
The variation of the structural features in the microstructure especially for material 326C,
resulted in that the etching method with nital and sodium metabisulfite provided best results
because it reveals most information of the microstructure, and etch lot of different features. The
etching method used for both trials are selected depending on the etchants ability to distinguish
martensite and non-martensitic transformation products. For Trial A Sodium metabisulfite
provided the best result.
49
a)
b)
Figure 35. a) From the work of Vander Vort a microstructure upper bainite (blue) and as-quenched martensite (brown)
etched with 10% sodium metabisulfite [44] b) material 326F etched first with nital and then with 5% sodium
metabisulfite, structure of lower bainite (blue) and as-quenched martensite (brown).
In order to get best result etching the samples with sodium metabisulfite, the following steps
should be considered:
1. 5g sodium metabisulfite powder and 100ml still water is mixed together, which provides
a solution of approximately 5% sodium metabisulfite. Use a fresh mix at every etching
occasion.
2. Etch a freshly polished surface in 2% Nital for 2-5s to remove possible oxide layer.
3. Etch the sample in the solution with 5% sodium metabisulfite in still-water in 30-60 s.
4. Put the sample in a crucible with water, do not rinse under running water. Rinsing under
running water could destroy the deposited sulfide film.
5. Dry the same.
In Trial B the samples were tempered, the sodium metabisulfite method did not work as good for
the tempered samples. Instead another method was used, consisting of picral and HCl. From the
work of C.D Liu et.al, [2], pictures of mixed microstructure of bainite and martensite was etched
with 4% picral and HCl. A comparison between their work and the result from Trial B where the
same etching method was used can be seen in Figure 36. Trial B was tempered in order to level
out the hardness profile and to produce tempered martensite. The picral and HCl solution showed
best result for Trial B.
50
a)
b)
Figure 36. Comparison between the work of, a) C.D Liu et.al, [2], and b) the result from Trial B sample 326F 10mm after
60s. Both samples is etched in 4% Picral and HCl.
One step in the production of injector bodies at Cummins Fuel Systems is a tempering step. The
tempering will result in tempered martensite. During the tempering fine dispersed carbides will
precipitate inside the martensitic phase. The diffusion of carbon during tempering at elevated
temperatures could result in a change of the martensite lattice parameters. This is further
discussed in section 2.4. Both tempered martensite and lower bainite have precipitated carbides.
But the distribution of carbides is different between the two structures. Even if the morphology
of the ferrite needles in a matrix of martensite reveal the lower bainite structures. It is necessary
to confirm the features in a SEM or TEM. The orientation of the precipitated carbides is different
for tempered martensite and lower bainite, this specific orientation can be studied in TEM.
51
6. Conclusions
326F material shows less amount of lower bainite structure.
Material 326F provides higher average surface hardness before tempering.
It is possible to create a variation of lower bainite content in a matrix of martensite by
irrupted isothermal heat treating in molten salt.
It is possible to distinguish bainte from martensite under LOM depending on the etching
methods.
Lower transformation temperature result in thinner needles and higher average hardness.
7. Further work
SEM or TEM to confirm bainitic and tempered martensitic constituents. Both strucutres
have characteristic carbide orientation within each phase distinguishing them from each
other.
A more quantitative study of heat treatments around MS to better understand the growth
mechanisms. Test more isothermal heat treating temperatures to better control the lower
bainite content.
More etching methods should be tested, to better understand etching parameters e.g. time
and concentration of chemicals.
In order to further investigate the bainite content, one way would be to study the
microstructure before tempering. As-quenched martensite does not have precipitated
carbides.
The work should be focused toward326 F material from Ovako AB, because it’s higher
hardenability, which seems to result in slower lower bainite transformation.
Perform XRD-analysis in order to see if any retained austenite is present in the
microstructure. Retained austenite has imparing effects on the mechanical properties.
EBSD- analysis to study the lower bainite refinement of the martensitic structure. This
study could provide a better understanding of the refining effects of the lower bainte
needles on the martensitic structure.
52
References
[1] Y. Tomita and K. Okabayashi, Heat treatment for improvememt in lower temperature
mechanical properties of 0.40 pct C-Cr-Mo Ultrahigh Strength Steel. Metallurgical
transactions V. 14A, Nov 1983 (2387-2393)
[2] C.D. Liu and P.W. Kao, Tensile properties of a 0.34C-3Ni-Cr-Mo-V steel with mixed lowerbainite-martensite structures, Material Science and Engineering A150 (1992) 171-177
[3] D.Y. Wei, J.L. Gu, H.S. Fang, B.Z. Bai, Z.G Yang, Fatigue behavior of 1500 MPa
bainite/martensite duplex-phase high strength steel, International Journal of Fatigue 26
(2004) 437-442
[4] J. Chakraborty, I. Manna, Development of ultrafine ferritic sheaves/plates in SAE 52100 steel
for enhancement of strength by controlled thermomechanical processing, Materials Science
and Engineering A 548 (2012) 33-42
[5] A. Abdollah-Zadeh*, A. Salemi, H. Assadi, Mechanical behavior of CrMo steel with
tempered martensite and ferrite-bainite-martesite microstructure, Materials Science and
Engineering A 483-484 (2008) 325-328
[6] George Krauss, Steel: Processes, Structure and Performance, ASM International, ISBN: 087170-817-5, August 2005
[7] G.E. Totten, G.E. Totten & Associates, LLC, M. Narazaki, Utsunomiya University (Japan),
R.R. Blackwood and L.M. Jarvis, Tenaxol Inc., Failures Related to Heat Treating Operations
[8] S. Morito, H. Yoshida, T.maki, X. Huang, Effect of block size on the strength of lath
martensite in low carbon steels, Materials science and engineering A438-440 (2006) 237-240
[9] S. Morito, X. Huang, T. Furuhara, T. Maki, N. Hansen, The morphology and
crystallography of lath martensite in alloy steels, Acta Materialia 54 (2006) 5323–
5331
Kangying Zhu, Olivier Bouaziz, Carla Oberbillig, Mingxin Huang, An approach
to define the effective lath size controlling yield strength of bainite, Materials Science
and Engineering A 527 (2010) 6614–6619
[10]
[11]
Parrish, Geoffrey, Carburizing Microstructures and Properties, ASM International,
December 1990
53
[12]
Peter Kolmskog Does Bainite form with or without diffusion? The experimental and
theoretical evidence, Kungliga Tekniska Hogskolan, ISBN: 978-91-7501-755-6, 24 may
2013
[13]
Zhi-Gang Yang , Hong-Sheng Fang, An overview on bainite formation in steels, Current
Opinion in Solid State and Materials Science 9 (2005) 277–286
[14]
H. K. D. H. Bhadeshia, Bainite in Steels - 2nd Edition, Published by the Institute of
Materials, Publication Date: 9 march 2001, IBSD: 1-86125-112-2
[15]
Mats Hillert, Paradigm shift for bainite, Scripta Materialia 47 (2002) 175–180
[16]
H. K. D. H. Bhadeshia, LOW–CARBON STEELS: THE CALCULATION OF MIXED–
MICROSTRUCTURES & THEIR MECHANICAL PROPERTIES, University of
Cambridge/Research and Development Corporation of Japan Department of Materials
Science and Metallurgy Pembroke St., Cambridge CB2 3QZ, U. K.
[17]
Khodamorad Abbaszadeh, Hassan Sagha¯ any and Shahram Kheirandish, E®ect of
Bainite Morphology on Mechanical Properties of the Mixed Bainite-martensite
Microstructure in D6AC Steel, J. Mater. Sci. Technol., 2012, 28(4), 336{342
[18]
Z. Ławrynowicz, Transition from upper to lower bainite in Fe – C – Cr steel, Materials
Science and Technology November 2004 Vol. 20
[19]
J.W. Christian, The Theory of Transformations in Metals and Alloys (Part I + II),
Pergamon December 2002
[20]
Yoshiyuki Tomita, Review: Morphology control of ductile second phase and improved
mechanical properties in high strength low alloy steels with mixed structure, Journal of
materials science 27 (1992) 1705-1715
[21]
Yoshiyuki Tomita and Kunio Okabayashi, Mechanical Propeties of 0.40 Pct C-Ni-Cr-Mo
High Strength Steel Having a Mixed Structure of Martensite and Bainite, Metal. Trans. A
Volume 16A, January 1985 (73-82)
[22]
C.H young and H. K. D. H. Bhadeshia, Strength of mixtures of bainite and martensite,
Mater. Sci. Technol. 10 (1994) 209-214
[23]
George E. Totten, Steel Heat Treatment, Metallurgy and Technologies, Chapter 9
Quenching and Quenching, Technology, by Hans M. Tensi, Anton Stich, and George E.
Totten
54
[24]
H.-G. Lambers, S. Tschumak, H.J. Maier, D. Canadinc, On the bainitic and martensitic
phase transformation behavior and the mechanical properties of low alloy 51CrV4 steel,
International Journal Of Structural Changes In Solids, 3(2011) 15-27
[25]
H. Ohtani, S. Okaguchi, Y. Fujishiro and Y. Ohmori, Morghology and Properties of Lowcarbon Bainite, Metal. Trans. A Volume 21A 1990 (877-888)
Kh. Abbaszadeh, Sh. Kheirandish, H. Saghafian, The Effect of Lower Bainite
Volume Fraction on Tensile and Impact Propertiesof D6AC Medium Carbon Low
Alloy Ultrahigh Strength Steel, Iranian Journal of Materials Science & Engineering
Vol. 7, Number 3, Summer 2010
[26]
[27]
RICHARD FARRARA, Comparison of isothermal salt quench and oil quench plus
temper of 4XXX low alloy steels, US Army Armament research, Development and
Engineering Center close Combat armaments center Benet Laboratories Watervliet, N.Y.
12189-4050, Feb1996
[28]
A.H. Meysami, R. Ghasemzadeh , S.H. Seyedein, M.R. Aboutalebi, An investigation on
the microstructure and mechanical propertiesof direct-quenched and tempered AISI 4140
steel, Materials and Design 31 (2010) 1570–1575
[29]
G.R. Speich and W.C. Leslie ,Tempering of steel,Metallurgical transactions volume 3,
p.1043-1054, May 1972
[30]
B.L. Bramfitt and J.G. Speer. A perspective on the morphology of Bainite, , Metallurgical
Transactions A volume 21A p. 817-829, 1990 April.
V. García Navas, O. Gonzalo, I. Quintana, T. Pirling, Residual stresses and
structural changes generated at different steps of the manufacturing of gears: Effect
of banded structures, Material Science and Engineering A 528(2011) 5146-5157
[31]
[32]
http://books.google.se/books?id=ibg3GWDxFRsC&lpg=PA28&ots=YsCaR7Gzmh&dq=
segregation%20soaking%20homogenization&hl=sv&pg=PA34#v=onepage&q=segregation
%20soaking%20homogenization&f=false, 9:23 AM 4/28/2014
[33]
ASM International TTT, http://www.wpi.edu/Pubs/E-project/Available/E-project020210-161513/unrestricted/JOMINY_END_QUENCHING_OF_4140_STEEL.2.pdf
[34]
Daniel H. Herring, Oil Quenching,
http://www.heattreatdoctor.com/documents/Vacuum%20Oil%20Quenching%20Tech.pdf ,
3/3/2014 3:05 pm
[35]
Gajen P. Dubal, Salt bath quench, Heatbath/Park metallurgical group, Detroit Michigan
55
[36]
S. Raygan, J. Rassizadehghani, and M. Askari, Comparison of Microstructure and
Surface Properties of AISI 1045 Steel After Quenching in Hot Alkaline Salt Bath and Oil,
168—Volume 18(2) March 2009 Journal of Materials Engineering and Performance
[37]
Justin Lefevre and Kathy L. Hayrynen, Austempered Materials for Powertrain
Applications, Journal of Materials Engineering and Performance,V olume 22(7) July 2013
(1914–1922)
[38]
S.M.C. Van Bohemen, M.J. Santofimia and J. Sietsma, Experimental evidence for
bainite formation below MS In Fe-0.66C, Scripta Materialia 58 (2008) 488-491
[39]
http://en.wikipedia.org/wiki/41xx_steel, 09:43 am, 3-3-2014
[40]
Harry Chandler (editor), Heat theater’s guide: Practices and procedures for Irons and steels, ASM International ISBN 0-87170-520-6 (p. 319-325)
[41]
Jospeh R. Davis, ASM Handbook: Volume 4 Heat Treating, ISBN 0-87170-379-3
Bruce L.Bramfitt, Arlan O.Benscoter ,Metallographer's Guide: Practice and
Procedures for Irons and Steels, ASM International, March 2002
[42]
[43]
John E. Hilliard John W. Cahn, An Evaluation of Procedures in Quantitative
Metallography for Volume-Fraction Analysis, TRANSACTIONS OF THE
METALLURGICAL SOCIETY OF AIME, VOLUME 221, APRIL 1961 (344-352)
[44]
http://www.georgevandervoort.com/metallography/general/iron-and-steel/20001270microstructure-of-ferrous-alloys.html, 1:34 PM 29/04/2014.
56
Was this manual useful for you? yes no
Thank you for your participation!

* Your assessment is very important for improving the work of artificial intelligence, which forms the content of this project

Download PDF

advertisement