Materials Technology - Microalloyed Steels

Niobium Carbide Precipitation in Microalloyed Steel

Christian Klinkenberg*, Klaus Hulka*, Wolfgang Bleck**

• Niobium Products Company GmbH, Dusseldorf/Germany; •• Institute for Ferrous Metallurgy, RWTH Aachen University, Aachen/Germany.

The precipitation of niobium carbo-nitrides in the austenite phase, interphase and ferrite phase of microalloyed steel was assessed by a crit-
icalliterature review and a round table discussion. This work analyses the contribution of niobium carbide precipitates formed in ferrite in the
precipitation hardening of commercially hot rolled strip. Thermodynamics and kinetics of niobium carbo-nitride precipitation as well as the ef-
fect of deformation and temperature on the precipitation kinetics are discussed in various examples to determine the amount of niobium in
solid solution that will be available for precipitation hardening after thermomechanical rolling in the austenite phase and successive phase
transformation.

Keywords: microalloyed steel, niobium precipitates, recrystallisation, precipitation hardening, coherent and incoherent precipitates

Introduction allowing a strength increase by precipitation hardening in

the ferrite phase.
Niobium microalloyed high strength low alloy (HSLA)
steels are widely used in civil construction, automobile and Thermodynamics of NbC precipitation in
line pipe applications. These steels rely on thermomechani- austenite
cal rolling, a technique that simultaneously provides high
strength and toughness by grain refinement. Since these After a critical analysis of the literature concerning the
steels typically contain low carbon contents, they also have equilibrium solubilities in nearly stoichiometric composi-
excellent welding and cold forming properties. tions of pure austenitic Fe-Nb-C or Fe-Nb-N alloys [2-5],
The most important role of niobium as a microalloying Kunze proposed the following equations:
element in thermomechanically rolled steel is the retarda-
tion of austenite recrystallization, which provides more nu-
-Fe 7407
clei for the I/O: transformation and thus a finer grain size. 10gK~bc = 10g([%Nb][%C]) = -T +2.783, (1)
Besides niobium's role in solid solution by delaying all dif-
fusion controlled processes, its tendency to form carbides
y-Fe 9940
provides the dominant effect [1] (figure 1). Depending on 10gK NbN = 10g([%Nb][%N]) = -T + 3.82 (2)
the hot rolling parameters, such as deformation amount and
rate, temperature and interpass time, niobium carbide pre-
cipitation is generally incomplete with regard to the ther- where K is the equilibrium constant and T is temperature in
modynamic equilibrium state. As a result, a part of the nio- Kelvin. Since usually both elements carbon and nitrogen
bium stays in solid solution after finish rolling in the austen- are present in the steel, in most cases niobium forms carbo-
ite phase and is effective in retarding the transformation or nitrides. Taking into consideration the strong interaction be-
tween niobium and carbon by means of the interaction pa-
rameters e( and rF, the solubility of niobium carbo-nitride
can be described by:
100
y)
10g([%N b][%CP[%N]l-
80 -YI----l------t0.08 ~
~
c
= Y .10gK Nbc + (1- y) .10gKNbN
+ Y . log y + (1 - y) . log (1 - y) - (e~b + e~) . [%C]
-
0

0
0::
0)
60
- r~~ . [%Cf (3)
.!:: 40
c
-
C1J

0
Vl
"0
~ C
where eNb + ec =
C 2160
-T - cc 987
0.565, rNb = T - 0.435,
20 -r--r--10.02 0
c. and y is the carbon fraction in the preciptiates. Applying
·u
....
C1J
equation (3) to the example of a steel containing 0.03 % Nb,
o (afterYamamoto) 0
a...
0.05 % C and 0.01 % N (always weight%) with y = 0.5 and
10-1 10 0 10 1 10 2 10 3 10 4 T = 950°C results in an equilibrium solute concentration of
Holding Time in s 0.007 % niobium available for precipitation in the ferrite.
This calculation makes clear that 77 % of the initial niobi-
Figure 1. Influence of dissolved and precipitated niobium on the re- um content is already precipitated during austenite process-
tardation of austenite recrystallisation [1]. ing.

744 steel research int. 75 (2004) No. 11

 Materials Technology - Microalloyed Steels

For a more practical approach

Mn - Nb - Steel Mn-Nb- Ti - Steel
another solubility product for nio-
NbN NbC
bium carbo-nitride is commonly 25
used [6]: c
~
OJ

CJ
20
Cl.

log [(%Nb)(C + 12/14 N)] = c

0 TiN TiC
u 15
6770 CJ
~

--+2.26 (4)
T ""CJ

'0 10
~

Even though in HSLA steels OJ

.s::::>
E
mainly niobium carbo-nitrides are ~
5 f---+---+---+--V
observed, pure niobium carbides =
can form when the nitrogen is
fixed by titanium additions. 0.43 0.44
Figure 2 gives an example of a
0.08% C, 1.6% Mn, 0.04% Nb lattice constant in nm lattice constant in nm
steel with an addition of 0.02% Ti
Figure 2. Frequency distributions of lattice constants in TM-treated heavy plate steel (0.08% C,
[7]. 1.6% Mn, 0.04% Nb + 0.02% Ti) [7).
Figure 3 shows the tempera-
tures required for complete disso-
lution of the carbides and carbonitrides of niobium that 0.20
have formed during the cooling of cast slab [2-6]. It is ob- - Niobiumcarbide
vious that even with a traditional HSLA steel of 0.10 %C ~ log [Nb)[C] = 2.78-7407fT
• 0.16
and 0.03 %Nb the niobium compounds will be completely
dissolved during reheating to the hot rolling temperature,
...s
.5
~
- - Niobiumcarbonitride
0.12 log [Nb)[C+12114N) =
which typically is 1200°C and more. Much higher niobium ~
0
contents can be utilised at a lower carbon level or if nitro- u 2.26-6770fT
~
gen is fixed by titanium. z 0.08
CD
:a:::s .-
On the other hand, the niobium carbide volume fraction,
available for precipitation after complete dissolution and '0 0.04
,.. ~

........
en
-
".
austenite processing, is schematically represented in 0.1% [C+12114N)
figure 4 [8]. It is obvious that, according to equilibrium 0.00
considerations, the percentage of niobium not precipitated 800 900 1000 1100 1200 1300
in austenite is higher the closer the chemical composition of
Soaking temperature in ·C
the steel is to the stoichiometric ratio of carbon to niobium.
Thus the amount of solute niobium available for precipita-
Figure 3. Solubility of niobium carbide and niobium carbonitride in
tion in ferrite becomes greater with reduced carbon in
the austenite phase of low-carbon steel [2-6).
HSLA steels.

Precipitation during austenite processing

Log ([%Nb) [%C)) = 2.78 -7407fT

Cooling down to the rolling temperature leads to a niobi-

um-supersaturated austenitic matrix. Niobium carbide par- 0.06 f------1f----+------+------+---+----;
ticles can nucleate homogeneously inside the matrix or,
more likely, heterogeneously on dislocations, subgrain,
grain and phase boundaries as well as on other precipitates Steel composition
~0.04
like aluminum or titanium nitrides. Deformation can initiate c
o
spontaneous precipitation while kinetic effects delay static u
c:
formation of niobium carbide in the undeformed austenite o

to several minutes. A schematic representation of the pre- ~ 0.02~~~~~~

cipitation start and finish temperatures in undeformed and
strained austenite is given in figure 5. In a steel containing
0.03 % Nb and 0.05 % C + O.oI % N "static" formation of
niobium carbides has its maximum nucleation rate after o 002 004 0.06 0.08 0.10
appro 100 seconds at about 950°C. "Dynamic" precipita- Niobium content in %
tion, however, starts without a long incubation time while Figure 4. Schematic demonstration of NbC volume fraction avail-
the maximum nucleation rate shifts to the lower tempera- able for ferrite precipitation as a result of the steel's carbon content
ture of about 900 °C [9]. [8).

steel research int. 75 (2004) No. 11 745

Materials Technology - Microalloyed Steels

Several models have been developed describing the pre- tion temperature of 1050°C the driving force for re-
cipitation kinetics of niobium carbonitrides in deformed crystallization becomes so strong that recrystallization
austenite. That of Dutta and Sellars [10] is based on classi- takes place before the onset of precipitation. In this case the
cal nucleation theory regarding the deformation rate and the precipitation kinetics are not promoted by deformation.
Zener-Hollomon parameter. Improvements can be made by Figure 7 shows the absolute value of niobium in solid so-
considering excess vacancies which are present during de- lution as a function of finish rolling temperature for two dif-
formation. According to Liu [11] relatively stable niobium- ferent pipe steels which were industrially produced apply-
vacancy pairs with a high diffusivity are formed which de- ing the typical thermomechanical rolling schedules for a
compose preferably at dislocation walls. The diffusion of plate and a hot strip mill, respectively [8]. This result was
niobium atoms is a much slower process. Niobium carbides achieved by means of a chemical extraction technique and
and carbo-nitrides nucleate along dislocation lines or dislo- provides quantitative confirmation that the higher deforma-
cation nodes. Consequently, deformation-induced nucle- tion speed and the shorter interpass time in the finishing
ation takes place preferably along dislocation walls where stand of a hot strip mill do not allow precipitation in the
the local concentration product and the temperature deter- austenite to approach equilibrium.
mine the degree of precipitation.
The effect of deformation and temperature on the precip- Determination and characteristics of precipi-
itation kinetics is given in figure 6 [12]. During deforma-
tates formed in austenite
tion in the lower austenite region niobium in solid solution
retards the onset of the diffusion-controlled recrystallization
The chemical extraction method is effective in determin-
process. Precipitation thus takes place on dislocations that ing the volume fraction of precipitates in the austenite. Af-
have been created during deformation. At a high deforma-
ter dissolving the steel specimen, precipitates remain in the
filter residue. Analysis of the filter residue not only deter-
mines the amount of precipitated versus total alloy content
but also allows determination of the precipitate's constitu-
1050r----.--~--_r_-.....,.-.,..,.....-....,....,~_:_:_-_,
static precipitation tion by X-ray diffraction studies.
(undeformed Austenite)
For nanometer-sized precipitates, transmission electron
dynamic precipitation )h microscopy (TEM) is the most powerful means to identify
'-' 1000 (during 10-3 s-1 deformation ) ~'I-+-----l
o their presence, analyse their chemistry and study their loca-
c:
tion at lattice-defects like grain boundaries and dislocations,
~
or in the host lattice in the case of bulk precipitates. Special
::J 95011----+_#..
E
QI
emphasis has to be given to sample preparation. Bulk sam-
Cl.
E ples are prepared by electrochemical or ion beam thinning
QI
>- and precipitates can be isolated by the carbon extraction
90011-----1""
replica method.
With a high resolution TEM precipitates of about 1-2 nm
850L-_..J.:!..._--L_---J_......::!...l..-_.......L..._......J..:------J and information on the coherency of bulk precipitates and
10-1 10 10 2 possible interfacial dislocation structures can be obtained.
Time in s Analytical investigations can be performed by energy dis-
persive x-ray spectroscopy (EDS), electron energy loss
Figure 5. Precipitation of Nb(C, N) in undeformed or slow deformed spectroscopy (EELS) and more recently also new tech-
austenite [9). niques have been offered by energy filtering TEM
(EFTEM). From the projected particle distribution revealed
in TEM micrographs, the particle volume fractions can be
100
determined by normalising with the foil thickness, which
Def. Temperature [0C]
can for example be measured by counting thickness fringes
e 900 950 1050
0.00 o ~ 0 under two-beam contrast conditions [13]. The typical size
~
75
0.44 ••• fractions for titanium and niobium containing carbonitrides
in industrially produced hot strip material are given in
J:l Table 1 [13]. The examined hot strip material was hot
'ii, 50
.~ rolled after reheating at 1230°C with a roughing tempera-
a. ture of 1160°C and finish rolling start and finish tempera-
.c
z tures of 1100 and 870°C. It underwent accelerated cooling
25
at 20 K S-1 to the coiling temperature of 630°C.
In another study [14] the particles that did not dissolve
0 during chemical extraction were identified as TiN particles
100 101 102 103 100 similar to those already existing in the slab. These particles
Time [s] had a certain effect on austenite grain size control as they
were stable during reheating prior to hot rolling. On the oth-
Figure 6. Effect of deformation and temperature on the precipitation
kinetics of Nb-carbide [12).
er hand the eutectic particles were dissolved at reheating

746 steel research into 75 (2004) No. 11

 Materials Technology - Microalloyed Steels

temperatures above 1050 "C. The niobium in solid solution

a - Fe 11030
was re-precipitated in the form of strain induced 20 to 30 1og K NbC = - - - +4.90 (5)
T
nm NbC particles during austenite processing. The chemi-
cal composition of the precipitates in the filtration residue
a - Fe 12000
indicates that the coarser particles are close to pure TiN, 1og K NbN = - - - +4.93 (6)
T
while the finer particles are almost pure NbC [8] (figure 8).
All particles were said to be incoherent. Carbides and ni-
trides of niobium and titanium exhibit a face centered cubic It can be seen in figure 9 that for a steel containing
structure and their carbo-nitrides may appear in complex 0.03 % Nb, 0.05% C and 0.01% N about I ppm Nb will re-
form. main in solid solution when thermal equilibrium is reached
at 800°C [15].
Thermodynamics of niobium carbonitride
precipitation in ferrite Precipitation kinetics during cooling

The niobium in solid solution at the finish-rolling tem- Solute niobium, which has not been precipitated during
perature is available for the formation of fine niobium car- austenite processing, may form niobium carbides in the in-
bo-nitride precipitates in ferrite. These are of the appropri- terphase between austenite and ferrite as well as inside the
ate size for providing a strength increase via precipitation ferrite grains.
hardening. A process, analogous to niobium carbo-nitride
formation in austenite, has to be expected for precipitation
in ferrite. The equilibrium solubility in ferrite can be de-
duced from the solubility products valid for austenite. Carbide Nitride

0.4501-----+--+--+--1-------1

0.10%t.1.50%....0.007%N.0.04%1Il.0.04"1oV 0.03'.t.180"loMn.OOO4%N.Ol0"loNb.0015'.Ti
.
se 0.06 I

,
E
c - Equilibrium condition p/ c:

.....
;: 0.05 (1) PIl(C. N]: 1og([NbIC+12114N]) =2.26-6nOfT

~ 0.04
(2) NbC : 1og([NbIC» = 2.78-7407fT
...... Plole production
c:
~
Q;
E
eo
0.440

j,r
c
0 - 0 Strip production 0-
Go
g 0.03 J 0>
.~
u
E
1;1 !J'/ V 0.430 1--~~...t---+--1--
~ 0.02 --'

--d V I.A"" V
o
'c (~

..,..-- ~ --...
~ 0.01
~ V
::I
"0
V) 0
l,......--'
~ V
750 SOD 850 900 950 1000 750 800 850 900 950 1000 0.42 0o!=----:!:::--~:l::_-_:!:_-_:l:_-....,_!
Finish rolling temperoture (FRll in Dt 20 40 60 80 100
Nitrogen content in %

Figure 7. Solute niobium at finish rolling temperature for two pipe steel alloy concepts [8]. Figure 8. X-ray lattice parameter of carboni-
trides in extracted residue [8].

Table 1. Characteristics of niobium carbonitride precipitates in an industrially hot rolled 0.06%C, 0.6% Mn, 0.006% N Steel with 0.02% Nb
and 0.06% Ti [13].

Type of precipitate Statistics of Mean particle Particle volume % of microalloy

observation size [nml fraction addition
Eutectic >300 particles in >1000 5.6xlO-4 42.7
carbonitrides -10!-lm2 area
Undissolved 26 particles in 67 0.8xlO-4 6.1
carbonitrides -500!-lm2 area
Fine carbonitrides - 2500 particles 6.8 3.6xlO-4 27.5
after rolling in -1O!-lm2 area
~ = 76.3%

steel research int. 75 (2004) No. 11 747

Materials Technology - Mlcroalloyed Steels

An atomistic Monte Carlo method for simulation of the ditions. This formation path can be explained by the com-
homogeneous nucleation and growth of niobium carbide petition between the driving force for NbC precipitation and
precipitates in ferrite was proposed [16,17]. This model the fast diffusion of interstitial carbon, which compensates
uses a vacancy diffusion mechanism for iron and niobium the weak driving force for iron carbide precipitation at a
atoms and an interstitial mechanism for the carbon atoms. temperature from 5oo to 700°C.
Besides the expected homogeneous niobium carbide pre- The traditional method to demonstrate the effect of pre-
cipitation mechanism, the simulation exhibited unexpected cipitation in ferrite was the solution treatment of niobium
formation of transient iron carbide clusters prior to the NbC containing steel (e.g. at 1250 "C) followed by water
precipitation for certain temperature and concentration con- quenching, in order to avoid any precipitation in the austen-
8-y------------,------, ite. The as-treated samples were isothermally annealed at
various temperatures in the ferrite region and the hardness
E 7 - Niobiumcarbide development was measured over the annealing time.
CL
log [Nb)[C] = 4.90-11030IT
.~6 --+-I,-:-:c=-:--=--~
0.05% C
Figure 10 shows such a result [18] indicating that there is a
1: 5 maximum hardness increase when annealing time and tem-
- •Niobiumnitride
f log [Nb)[N] = 4.93-12000IT -+--+-.r-------f
perature are optimised. This maximum hardness is a result
~ 4 of two influencin factors: volume fraction, which increases
~3+_-------____:____:____:__+__f_!_-----_i with longer annealing time and also particle size, which be-
CD
comes too large and thus less effective, when 'overaged'.
i2+----------~~'-T--------1
However, industrial cooling conditions are different. Af-
~1+--------____::~A,..L_-------__1
ter finish rolling in the lower austenite region, the steel ex-
O+---~ ...~~-__---r"-...,....~ periences either air cooling (cooling rate about 0.5-1 °C/s)
600 700 800 900 1000 or accelerated cooling, which is interrupted at temperatures
Soaking temperature In ·C ranging from 650°C to 550°C. In this latter process the
cooling rate differs from about 10°C/s for 20 mm plates to
Figure 9. Solubility of niobium carbide and niobium nitride in the fer- almost 100 °C/s for thin hot strip which is finally coiled
rite phase of low-carbon steel [15].
(cooling rate about 0.5 DC/min).
Figure 11 [19] shows that the amount of incoherent pre-
or> 80....-----,.------,.------,....------, cipitates, determined by chemical extraction method, in-
>-
::c
<I creases when slow cooling is applied after finish rolling.
.s 60f--....,.'::"""O~+-___..F_=l.l~i:::oo'-----...J'l'---=-_--j
c: 500 0C
These rather big precipitates have only a minor influence
o
on the strength increase via precipitation hardening and
~
o
>
they can be promoted by either the growth of existing
~ 20t-+____::"-lJ'''---+-----t'''''ooo;:::-~---tOo_=_~---j austenite precipitates or the formation of interphase precip-
c: 5500C
"E itates. The same effect can be observed when the accelerat-
o
::c ed cooling is interrupted at a relatively high temperature
10 1 10 2
Annealing time in min followed by the slow cooling rate typical for a hot strip coil.
Figure 12 [12] gives the influence of holding temperature
Figure 10. Hardness of a 0.11 % Nb steel after annealing at various and isothermal holding time on the amount of incoherent
temperatures [18]. niobium carbonitride precipitates, underlining the former
statement. At the higher coiling temperatures of 700 to
50 800°C precipitation started prior to the onset of the austen-
ite to ferrite transformation and ended before the phase
CT in °c
I I transformation was completed. Thus only a small amount of
40 precipitates are formed after the austenite to ferrite trans-
700 t:.--- formation was completed. For all temperatures, except
600 0 -

...
c
~

.g2
~~,
h 500
25 .-_.-
~_._.-

400 e------
400°C where precipitation is inhibited, the major volume
fraction of the precipitates is formed during the austenite to
ferrite phase transformation. This does not necessarily im-
ply that precipitation takes place at the solute-enriched in-
u
.!:
.c
a IS\. eo,
\ \ .......
" , '- --- -- r--_.!~'
terphase boundary. It could also happen on existing austen-
ite precipitates or even inside the newly formed ferrite
'.-, e
;z:

a \ \\

o
, \

'- ~_:'..::..
e
:-M!:;;;;' --,--- y-
----iw:~
grains due to the lower solubility of niobium in ferrite than
in austenite. Figure 13 gives the strength increase for this
same steel as a function of cooling rate and coiling temper-
a 20 40 60 80 lOa ature. At a high cooling rate a large amount of niobium is in
Cooling rote in oC.s-1 solid solution and is available for precipitation in the ferrite.
Thus the maximum strength at a coiling temperature of
Figure 11. Influence of cooling rate on formation of incoherent Nb- 600°C is attributed to such precipitation of niobium car-
carbides [19]. bides, which are considered to be coherent.

748 steel research int. 75 (2004) No. 11

 Materials Technology - Mlcroalloyed Steels

The transmission electron microscopic (TEM) investiga-
tion of two steels having 0.07%C, 0.02/0.028%Nb and
0.33/1.1 % Mn revealed fine (3.6-7.2 nm) niobium carbide
precipitates that formed in austenite [21]. These strain-in-
duced precipitates that have formed on austenite deforma-
tion bands are typically found in high manganese samples
that were air cooled to room temperature. In addition, these
samples showed a small number of interphase and ferrite
precipitates. 68 to 90% of the initial niobium content was
found in the precipitates of these samples. Accelerated cool-
ing (50°C/s to coiling temperatures of 650 and 680°C re-
spectively) from a finish-rolling temperature of 892°C sup-
pressed interphase and ferrite precipitation and only 43 to
47% of the initial niobium content was found to be precip-
itated.
A second transmission electron microscope based inves-
tigation came to a similar conclusion, finding that there is
no interphase or ferrite precipitation after fast cooling at
20°C/s to coiling temperatures between 630 and 670°C
[13]. These investigations were carried out on industrially-
produced hot strip having 0.05%C, 0.02%Nb, 0.06% Ti and
0.55% Mn. The amount of precipitated niobium was about
50% of the initial niobium content. Further examination of
as-coiled steels having a similar chemical composition but
a cooling rate of only lO°C/s showed precipitation on the
deformation-induced austenitic dislocation substructure.
Carbonitride precipitation in austenite was identified by the
electron diffraction method in all examined grains. While
interphase precipitates were detected in some few grains,
ferrite precipitation seemed to be completely suppressed in
these samples [22].
Morphology of precipitates formed in ferrite
tionship between the particles and the ferrite matrix. This
aspect of interphase precipitation has frequently been mis-
interpreted as random carbonitride precipitation in supersat-
urated ferrite [35].
Some authors have considered coherent or semi-coherent
NbC particles [8,19,25-28]. Coherent precipitates can be vi-
sualised in transmission electron microscopy by the strain
field in the surrounding matrix caused by the lattice mis-
match of the particles with respect to the matrix. The strain
field can be seen in images obtained with bright or dark
field techniques. However experimental evidence has so far
not been reported in the literature.
Niobium carbonitride precipitates that formed in the fer-
rite phase of the HSLA (Nb) hot strip given in table 3, were
reported to be semi-coherent. The strip was finish rolled at
900°C and high temperature coiled at 680°C. The particle
size was about 10 nm in diameter. Nucleation took place in-
side the grains but sometimes also in the austenite-ferrite in-
terphase leading to a characteristic banded microstructure
[26].
The same result was reported for an IF (NbTi) steel grade,
exhibiting 30 nm diameter semi-coherent precipitates after
920°C finish rolling and high temperature coiling at 730°C.
In contrast, the ULC steel grades did not show any co-
;;-e 40
M • BOOoC
s=
I
• 700°C
.s 30 • 600°C
c: • 400°C
C
Cl
•
~20 a/y start (.5 '!o)
•
•+/
u
~
C1.

~10
Carbonitrides formed in the austenite phase are incoher-
LJ
•__ • •
ent and have an arbitrary orientation relationship to the sur-
.s::>
:z:
u
of:
•
L

rounding ferrite matrix of the final hot strip. In contrast, car- 10.0 100.0 1000.0
Coiling time in s
bonitride particles which have nucleated during or after the
y-a transformation have the Baker-Nutting orientation re- Figure 12. Influence of holding temperature and time on the amount
lationship to the surrounding ferrite matrix [23]. According of incoherent niobium-carbonitride precipitates [12].
to this relationship, either a <100> or <0 II > direction of the
carbonitride particle is parallel to a < I00> direction of the
ferrite unit cell. The correlations for the precipitates and the
III
ferritic matrix, given in table 2, correspond to these orien-
tation relationships [24]. The crystallographic relationship
between the orientation of the ferritic matrix and the parti-
cles can be determined by electron diffraction techniques
and the particle origin (nucleation in austenite or ferrite) is o
c...
established by comparison to the Baker-Nutting-relation- ~

ship [13,23]. .= 500

Interphase precipitates are formed in the solute-enriched
phase boundary between austenite and ferrite. These pre-
cipitates are arranged in sheets, parallel to the instantaneous
positions of the 'Y/ (): interface during transformation. When ~~----
oriented nearly parallel to the incident electron beam, these
sheets exhibit a line-like arrangement ("precipitate row for- 4001-.----L.------'-------'-----' 125
mation"). For all other orientations, which are the large ma- o 200 400 600
Coiling temperature in °C
jority, interphase precipitates appear as a random particle
distribution and can only be identified by means of electron Figure 13. Strength increase as a function of cooling rate and coil-
diffraction analysis of the crystallographic orientation rela- ing temperature [12] (T.S.: tensile strength).

steel research int. 75 (2004) No. 11 749

Materials Technology - Microalloyed Steels

Table 2. Lattice mismatch for NbCxNy precipitates in austenite and tate volume fraction and particle size
on the yield strength increment calcu-
Requireddistortionof matrix [%] lated according to the Ashby-Orowan
Orientation
Matrix equation and taking into account ran-
relationship NbC NbCo.s NbNo.s dom distribution is given in figure 14
[100] ppt II [100] y 25.5 26.6 23.0 [31]. However, according to other
Y [010] ppt II [101] y 25.5 26.6 23.0 studies, it is possible to find different
results and there is at least a certain
[001] ppt II [001] Y 25.5 26.6 23.0
confusion about the quantitative
[100] ppt II [100] y 56.3 57.7 53.1 strengthening predictions related to
a [011] ppt II [010] a 10.5 11.5 8.4 the presence of niobium carbo-ni-
[011] ppt II [001] a 10.5 11.5 8.4 trides.
Using the Ashby-Orowan model it is
possible to calculate the hardening ef-
Table 3. Chemical compositions of niobium-treated steels [26-28) in weight%. fect from niobium in solid solution
[17]. The procedure involves a quench
hardening and ageing treatment. Car-
Steel C Mn Al N Nb Ti bon in solid solution is supposed to be
IF (NbTi) /27/ 0.0030 0.12 0.041 0.0020 0.025 0.019 entirely trapped by defects, and ferrite
ULC (Nb-Ti) /27/ 0.0032 0.18 0.045 0.0036 0.017 0.009 grains do not change in size during the
ULC (Nb) /27/ 0.0031 0.19 0.035 0.0021 0.019 - ageing. Consequently, these effects
cannot contribute to any hardness
HSLA (Nb) /26/ 0.0480 0.40 0.036 0.0048 0.027 0.001
change. The calibration of the model
X70 /8,36/ 0.030 0.18 0.030 0.004 0.10 0.015 leads to good agreement with the ex-
perimental results and allows the de-
termination of an effective contribu-
tion of niobium in solid solution to the
herency of the niobium carbo-nitride precipitates that had increase in flow stress equal to 0.24 MPa per weight ppm of
formed in the ferrite phase of the hot strip coiled at 740°C niobium. This very high value cannot be due to solid solu-
[27]. The chemical compositions of these steels are given in tion only, and is expected to originate from very small clus-
table 3. ters or aggregates of Nb and C present in the as-quench
As there exists a considerable lattice mismatch between state. Those clusters are assumed to be precursors of the
the niobium carbonitride precipitates and the ferrite matrix, precipitation of the equilibrium NaCl-type NbC particles.
some researchers claim that these particles are generally not DeArdo and Kestenbach did not find precipitates which
coherent with the ferrite matrix [13,29]. had formed in the ferrite in material processed according to
The industrially produced X70 steel given in table 3 re- industrial conditions [13,21,22,32-34]. In this case inter-
vealed incoherent Nb(CN) precipitates after finish rolling phase precipitation should offer the highest precipitation
below the 1/a-transformation temperature, i.e. below strengthening potential as it is creating the smallest parti-
no°e. Finish rolling in this temperature region was sup- cles. Strain induced precipitation in austenite might also
posed to generate significant dislocation hardening [8,36]. give a certain yield strength contribution, especially in hot
strip production where the short interpass time prevents ex-
Quantification of the strengthening effect of cessive growth [22]. According to [13] precipitation after
precipitates transformation to ferrite would preferably occur on existing
precipitate particles that have already formed in austenite
All strengthening mechanisms are based on the hamper- and thus have no significant strengthening effect. However,
ing of dislocation movement. Dislocations may cut through other researchers have observed strengthening by precipita-
coherent particles, making them weaker obstacles to dislo- tion in the ferrite during coil cooling of hot rolled strip af-
cation motion. However, as is known from other materials ter coiling between 500 and 600 °C [12].
systems, a high enough density of coherent precipitates may There have been various attempts to calculate the yield
also induce a significant strengthening effect, to which a strength increase by precipitation hardening (L\YS PPTN) in
number of different particle/dislocation interaction mecha- low carbon strip and plate steels.
nisms may contribute [30]. So it is not yet clear whether co- - Determination of precipitate volume fraction by TEM ob-
herent precipitates significantly contribute to strengthening servations and calculation of L\YSPPTN according to the
in microalloyed steels or not. Ashby-Orowan model [3,21,22].
Incoherent particles cannot be cut by dislocations and - Calculation by subtracting the measurable strengthening
have to be bypassed by means of the Orowan mechanism, effects from the observed yield stress (YSobserved) [20,29]:
resulting in a yield strength increase. The strength increase
by precipitation hardening of commercial microalloyed hot
rolled steels is affected by volume, size and distribution of L\YSPPTN = YSobserved-[YSO + L\YSSolid Solution +
the niobium carbonitride precipitates. The effect of precipi- L\YS Work hardening +L\ YSSubgrain]

750 steel research into 75 (2004) No. 11


Several results of both methods indicate that the precipi-
tation hardening effect might often have been overestimat-
ed in the past.
In general, a normalising treatment of thermomechanical-
ly rolled steels causes a yield strength decrease. This is due
to the reduction of the dislocation density, Ostwald ripen-
ing, grain coarsening and the change of coherent and semi-
coherent into incoherent and bigger precipitates [20,33,34].
The strength contribution of interphase precipitates has
been determined in a recent investigation [34]. Tensile and
microhardness tests have been carried out on a commercial
0.11%C, 1.54%Mn, 0.04%Nb and O.ll %Ti hot strip that
had been finish rolled at 985°C and cooled at lOoC/s to a
coiling temperature of 650°C. The authors observed a drop
in yield strength from 603 MPa for the as-coiled strip to 462
MPa after a 30-minute normalising treatment at 900°C. Af-
ter allowing for the effects of dislocation density, Ostwald
ripening and y-precipitates, the macroscopic strengthening
effect of NbC interphase precipitates was calculated to be
56 MPa. As interphase precipitation only occurred in about
one half of the ferrite grains, the strengthening effect must
have been 112 MPa in these grains.
Besides precipitation strengthening, interphase precipi-
tates may cause toughness degradation as a result of their
sheet-like arrangement. However, grain refinement by nio-
bium metallurgy provides outstanding toughness and
strength in HSLA steel grades. This toughness reserve per-
mits a further strength increase by precipitation hardening,
even if this strengthening mechanism implies some tough-
ness reduction. In the case of severe toughness require-
ments, precipitation hardening by interphase precipitates
might be avoided, e.g. by fast cooling in the two-phase tem-
perature range.
Summary and Conclusion
The established principles of niobium carbide precipita-
tion and dissolution in austenite permit the estimation of the
amount of free niobium available for creating precipitates in
the ferrite. Plate rolling with its long interpass times offers
ample opportunity for precipitation and Ostwald ripening.
Thus niobium carbides or carbonitrides that have already
formed in the austenite phase may become too big to give a
significant contribution to precipitation hardening. Howev-
er strain-induced precipitation in austenite supposedly con-
tributes to the yield strength in hot strip production, where
the short interpass time prevents excessive growth [22].
Niobium carbide precipitation in ferrite is observed in
niobium-bearing steels subjected to industrial as well as
laboratory thermal and thermomechanical processes. The
probability of forming niobium carbide precipitates in fer-
rite or in the austenite-ferrite interphase during phase trans-
formation is dependent on the finish rolling and cooling
conditions. Precipitates formed in ferrite are typically found
after slow cooling from a finish rolling temperature of about
900°C and isothermal holding at coiling temperatures equal
to or above 600°C or after high temperature coiling above
680°C and coil cooling at about 0.5°C/min [12,19,26,27].
Analysing these results, there is a high probability of form-
ing genuine ferrite precipitates when the cooling rate is high
steel research into 75 (2004) No. 11
Materials Technology - Mlcroalloyed Steels
1000 ,---------------------"1
5nm
Particle
diameter
5nm
10nm
SOnm
10 L-_---l_....J...--lJL-L-L....l.....I...u..._----'--L.._-'---'-...J.-........................
lo-t 0.001 0.01
Viilume fraction of particles
Figure 14. Observed increments of yield strength in microalloyed
steels compared with the predictions of the Ashby-Orowan equation
[31).
(e.g. 50°C/sec) and the isothermal holding temperature is
around 600°C, while air cooling promotes the formation of
interphase precipitates. Investigations applying industrial
processing cooling rates of more than 10°C/s from the fin-
ish rolling temperature to typical coiling temperatures in the
range between 550 and 650°C did not show ferrite precipi-
tates when typical coil cooling rates of 0.5°C/min or higher
were used [13,21,22,32,34].
In one investigation, ferrite precipitates were created in a
dislocation-rich ferritic matrix by finish rolling below the
1/ a-transformation temperature [8,36].
There is still a certain disagreement with regard to the
morphology of the precipitates. It seems that it is compli-
cated to distinguish between coherent and semi-coherent
particles. Additionally, TEM observation of the strain field,
which is a characteristic of coherent precipitates, seems to
be difficult, particularly when the particles are very fine in
the sub-nm range. These problems may lead to different in-
terpretations of the results. Since size, distribution and co-
herency of precipitates in the ferrite are an important feature
of the strength contribution, the exact characterization of
the precipitates is a primary requisite for predicting their
contribution to precipitation hardening.
The physical metallurgical aspects of the effect of niobi-
um-carbide precipitates formed in ferrite seem to be basi-
cally understood. However, since relevant details for com-
mercially production of HSLA plate and strip are often not
available, there is no definite answer to the question
whether precipitation of niobium carbides significantly
contributes to strength or not. One open aspect seems to be
the existence of particles in the nm range as well as the
question if they are coherent with the ferritic matrix or not
[37].
Acknowledgement
The role of niobium carbonitride precipitation in ferrite
was assessed at the occasion of a round table discussion,
751
Materials Technology - Microalloyed Steels
when about fifty researchers from Europe, Asia, North and
South America met at the VDEh in Dusseldorf in 2000. This
event had been organised by the Niobium Products Compa-
ny (NPC) on the occasion of its 25th anniversary and was
held under the chairmanship of Professor Dr.-Ing. Wolfgang
Bleck from the RWTH Aachen. The discussion was unpub-
lished but reference to several contributions made at the
event are included in this paper. The authors are indebted to
all participants of the round table discussion.
(A2004030; received on 2 April 2004,
in final form on 19 May 2004)
Contact: Dr.-1ng. Christian Klinkenberg
Niobium Products Company GmbH
Steinstr. 28
40210 Dusseldorf / Germany
References
[I) S. Yamamoto, Ch. Ouchi, T Osuka: Thermomechanical Processing of
Microalloyed Austenite, TMS, Warrendale (PA), 1982, pp. 613/39.
[2) T.H. lohansen, N. Christensen, B. Augland: Trans. Met. Soc. AIME,
239 (1967),1651/4.
[3) S. Koyama, T. Ishii, K. Narita: 1. lap. Inst. of Metals, 35 (1971),
1089/94.
[4) Y.K. Lakshmanan, 1.S. Kirkaldy: Metall. Trans.A, 15A (1984),541/4.
[5) K. Balasubramanian, A. Kroupa, 1.S. Kirkaldy: Metall. Trans.A, 23A
(1992), 1263/6.
[6) K.J. Irvine, EB. Pickering, T. Gladman: 1. Iron Steel Inst., 205 (1967),
161/82.
[7) M.K. Graf, H.G. Hillenbrand, P.A. Peters: "Accelerated Cooling of
Steels", TMS Warrendale (PA), 1986, pp. I65n9.
[8) K. Hulka, 1.M. Gray, E Heisterkamp: "High Temperature Thermome-
chanical Processing of Pipe-Steel", Proceedings Pipeline Technology,
ed.by R. Denys, Elsevier, 2000, pp. 291/306.
[9) K. Weiss, 1.1. lonas: Met. Trans ASM/AIME, 10 A (1979), 831/40.
[IO)B. Dutta, C.M. Sellars: Mater. Sci. Techno!., 3 (1987),197/206.
[II) W.J. Liu: Metall. Mater. Trans.A, 26A (1995), 1641/57.
[12)l.C. Herman, B. Donnay, V.Leroy: ISH International, 32 (1992), No.
6,779/85.
[13)1tman, K.R. Cardoso, H.-l. Kestenbach: Materials Sci. and Technolo-
gy, 13 (1997), No.1, 49/56.
[14)S.Y. Subramanian, "Analysis of Solidification and Precipitation Be-
havior of HTP Slab", Research report of Mc Master University,
Hamilton Ontario, Oct. 1985.
752
[15)1. Kunze, private communication
[16) D. Gendt et al.: "Monte Carlo simulation of NbC precipitation kinet-
ics in a-Fe", Proceedings of the COM2ooo conference in Toronto, Au-
gust 2000.
[17)P. Maugis et al.: "A Model for Niobium Carbonitride Precipitation in
Ferrite", Proc. Thermec'2oo3, Mat. Sci. Forum Vols. 426-432 (2003),
pp.1313/18.
[18)H.-D. Bartholot, H.-l. Engell, W. vor dem Esche, K. Kaup: Stahl u.
Eisen, 91 (1971), No.4, 204/20.
[19)V. Leroy, 1.C. Herman: Final report ECSC contract No 721O-MA/20I,
1996, EUR 15748 Fr.
[20) E.A. Kothe: communication during round table event.
[2I)V.Thillou, M.Hua, C.I.Garcia, C.Perdrix, A.J.DeArdo: Proceedings
41st MWSPConference, ISS, Vol. XXXVII, 1999, p. 471/8.
[22)S.S. Campos, 1. Gallego, E.V. Morales, H.-l. Kestenbach: HSLA
Steels '2000, Ed.: Liu Guoquan, Wang Fuming, Wang Zubin, Zhang
Hongato, Beijing, The Metallurgical Industry Press, 2000, pp. 629/34.
[23)A.T. Davenport, L.C. Brossard, R.E. Miner: 1. Met.,27 (1975), 21n.
[24)A.J. DeArdo, 1.M. Gray, L. Meyer: "Niobium", 1984, Warrendale, PA,
TMS-AIME, pp. 685n59.
[25) Giith, communication during round table event.
[26)1. Kruse, R. Kaspar, E. Anelli, P.E. Di Nunzio: "Development of high-
strength strip steels with an improved processing potential", Final re-
port ECSC contract No 721O-EC/I09, 110, 404, 2000, EUR 19385
EN.
[27) Pichler, H. Spindler, C. Van Durme, P.E. Di Nunzio, K. Spiradek, S.
Geisler, W. Bleck: "Precipitation behaviour in the ferrite range of soft
microalloyed ULC steel grades", Final report ECSC contract No
721O-EC/119, 120, 121,213,408,2001, EUR 19878 EN.
[28) H.P. Schmitz, L. Kestens, P.E. Di Nunzio, 1. Zaitegui Hernandez, I.
Gutierrez Sanz: "Texture control in cold rolled steel sheets for an op-
timized anisotropy", Final report ECSC contract No 721O-EC/214,
944,945,410,122,2001.
[29)A.J. DeArdo: communication during round table event.
[30)1. Mayer: communication during round table event.
[3I)T. Gladman, "The Physical Metallurgy of Microalloyed Steels", The
Institute of Materials, London, 1997, ISBN 0 910716 81 2.
[32)A.C. Kneissel, C.1. Garcia, A.J. DeArdo: HSLA Steels: Processing,
Properties and Applications, Ed.: G. Tither, Z. Shouhua, The Minerals,
Metals & Materials Society, 1992, pp. 99/105.
[33)S.S. Campos, E.Y. Morales, H.-1. Kestenbach: Metallurgical and Ma-
terials Transactions A, 32A (2001), No.5, 1245/48.
[34) S.S. Campos, E.V. Morales, H.-1. Kestenbach: "A quantitative study
of interphase precipitation in a commercial microalloyed steel", Proc.
Thermec'2oo3, Mat. Sci. Forum Vols. 426-432 (2003), pp. 1517/22.
[35)H.-l. Kestenbach: communication during round table event
[36) K.Hulka, 1.M.Gray, E Heisterkamp, Niobium Technical Report,
NbTR-16/90, ISSN 0101-5963.
[37)M. Beres, T.E. Weirich, K. Hulka, H.-l. Kirsch, 1. Mayer: Steel Re-
search Int., 75 (2004), No. 11,759-765.
steel research into 75 (2004) No. 11