<Title of Publication> <Edited by…>

Materials Science and Technology (MS&T), 2006

A New Method of Characterizing and Quantifying Complex Microstructures in

Steels
J. Wu, C. I. Garcia, M. Hua, W. Gao, K. Cho, and A. J. DeArdo

Basic Metals Processing Research Institute

Department of Materials Science and Engineering, University of Pittsburgh
Pittsburgh, PA 15261, USA

Keywords: EBSD, Image Quality, Steel, Ferrite, Microstructure, Dislocation Density

Abstract

Through the 1970s, high strength microalloyed steels hot rolled to strip and plate had microstructures of
ferrite reduced-pearlite, and with yield strengths essentially limited to the range of 350-420 MPa. Since
that period, the automotive, energy and civilian construction industries have demanded steels of higher
strengths, while maintaining acceptable levels of other properties such as weldability, toughness, and
formability. Since the late 1970s, it has been recognized that the strengthening mechanisms present in
ferrite-pearlite steels had reached their limit in terms of grain refinement and solute and precipitation
hardening; hence the barrier of 350-420MPa was a real restraint. At about the same time, it was also
recognized that higher strengths in as-processed steels could only be achieved through the use of ferrite
of lower temperature formation, i.e., non-polygonal, acicular, bainitic or martensitic ferrite, either as a
monolithic matrix microstructure or as a combination. This change in achievable microstructures was
abetted by interrupted accelerated cooling, either on the runout table of a strip mill or after the finishing
pass in a plate mill. Hence, high strength hot rolled or CRA/CGL processed steels, with strengths in
excess of 420 MPa, now exhibit these complex microstructures. It is not uncommon today for a 490
MPa yield strength steel to exhibit a microstructure comprised of several types of microsconstituents:
non-polygonal ferrite, bainite, martensite and perhaps retained austenite. Traditional metallographic
techniques are no longer capable of analyzing these complex microstructures, especially in a quantitative
fashion. The inability to characterize and quantify these complex microstructures means that the true
strengthening mechanisms operative in these steels may be incorrectly understood and evaluated. The
recent application of the quantitative analysis of the image quality of the Kikuchi pattern resulting from
EBSD (EBSD-IQ) has led to an innovative way to quantitatively analyze complex microstructures in the
higher strength steels. This paper will present the technique and show where it has been used
successfully in research studies involving a broad range of steels including HSLA, MA, DP, TRIP, IF
and heat treated steels.

Introduction

During the decade of the 1970s, the world was experiencing an energy shortage. That was also the
period when steel metallurgy experienced a dramatic change in the alloy design and processing of high
strength, hot rolled steels. Through the 1950s and 1960s, high strength steels of yield strength (YS)
levels in excess of 300 MPa were either heat treated, low alloy martensitic or bainitic steels or high
carbon, hot rolled, plain carbon ferrite-pearlite steels. In the late 1960s, low carbon steels containing
microalloying elements (MAE) such as Nb, Ti or V started to become popular.[1-6] For example, the
combination of MAE, controlled rolling, and air cooling resulted in strength levels of 350-400 MPa in
light plate intended for pipeline applications. These microstructures of polygonal ferrite and pearlite
were rather simple and easy to recognize and to quantitatively characterize. Since this ferrite was
formed at relatively high temperatures, these microstructures consisted of equiaxed polygonal ferrite of a
given size, bounded by high angle boundaries, containing a low dislocation density, perhaps MA
precipitation and alloying elements in solid solution.

The strength of these simple microstructures was rather easy to understand and the observed yield
strength could be expressed as an expanded Hall-Petch equation, as shown below:[7,8]
 y   0   i  k y / d (1)
where ky is the fitting parameter (a material constant, i.e., the Hall-Petch slope), σo also a materials
constant for the starting stress for dislocation movement, d the grain diameter, σy the yield stress, and σi
the other YS increments, including solid solution, dislocation, and texture.[9-11]

This approach led to later equations that explicitly related the yield strength to steel composition and
microstructure:[9]
YS ( MPa)  15.4[3.5  2.1(% Mn)  5.4(% Si)  23(% N f )  1.13( d 1 / 2 )] (2)
where Nf is free nitrogen and d the ferrite grain size in mm. Equations of this kind led to attempts to
model the mechanical properties over a vast range of composition, processing and microstructure.

In the early 1980s, however, the world was requiring even higher strength hot rolled steels, and this
requirement led to a change in the approach to the alloy design and processing. In plate product, it was
found that higher strengths and better resistance to cleavage fracture could be achieved through the
combination of controlled rolling and interrupted accelerated cooling (IAC). Of course, high strength
MA strip steels had been produced in hot strip mills, with IAC on the runout table, since the early 1970s.
These two IAC practices are, of course, different. IAC on a plate mill is normally accelerated cooled
from the finishing pass at about 10-15°C/sec to the water end temperature (WET) in the range of 600-
400°C, followed by air cooling to RT. On a strip mill runout table the IAC cooling rate is around 20-
30°C/sec, where the WET is essentially the coiling temperature, normally in the range of 600-500°C,
followed by in-coil cooling of about 30°C/hr. In both cases, ferrite is formed at lower WET, with higher
strengths. In what follows, IAC steels and microstructures will refer to practices where the WET is low.

In the mid-1980s, two landmark conferences were held on IAC.[1-6] What was clearly demonstrated in
this and earlier work, was the simple fact that yield strength in excess of approx 400MPa in reasonable
thicknesses and with good toughness could only be attained in MA steels with low transformation
temperatures that resulted from IAC with low WET. MA steels processed to exhibit yield strength of
420 or 490 MPa or above, no longer had microstructures consisting of mainly polygonal ferrite. The
MA steels processed with controlled rolling and IAC exhibited complex microstructures consisting of
mixtures of polygonal, non-polygonal, acicular, bainitic and martensitic ferrite. This evolution is nicely
captured in Figure 1 which shows the change in processing, structure and properties for linepipe steels
over a two decade period of time.[12] An example of this kind of complex microstructure taken from
simulated strip processing with a low coiling temperature is shown in Figure 2.[13]

Not only are these higher strength microstructures more complex, but also the understanding of the
origins of strength in these steels was also much more difficult, since the complex microstructures were
no longer able to be analyzed using the approaches that worked well for the earlier polygonal ferrite-
pearlite steels. For example, polygonal ferrite is bounded by high angle boundaries and can be properly
analyzed by the Hall-Petch equation, with a measured grain diameter d taken from an etched optical
micrograph and a ky variously published as 16.9-18.1 (MPa/mm -1/2). To this can be added additional
contributions from solutes and precipitates, as indicated in equations 1 and 2, above.
 Increasing
X90 grain refinement Bainite
& decreasing
transformation T
e
ad TM-MACOS
X80 Ferrite + Bainite
Gr
-
Ferrite + Pearlite TM-treated
PI
X70 (~15% Pearlite)
A

Ferrite + Pearlite Hot rolled and normalized

X60
(~30% Pearlite)

Toughness
Figure 1: Evolution of plate steel for large diameter linepipe:
microstructure and mechanical properties[6]

Figure 2. Optical micrograph of a HSLA steel hot band, with low reheat, high roughing,
low finish rolling and coiling temperatures. Etched with 2% Nital.[13]

However, the situation is very different for the non-polygonal, acicular, and bainitic ferrite that often
results from IAC where the structural units or grains are separated by low angle boundaries or sub-
boundaries, even though some boundaries are often distinguishable and measurable with optical or SEM
microscopy and can yield an apparent “grain size”. In this case, the application of the conventional H-P
relation with the same high ky valid for high angle boundaries could vastly over estimate the grain
boundary or grain size contribution to strength for low temperature ferrite.

The absence of any reasonable method of validly analyzing the microstructure has led to
misunderstandings of the origins of strength in these complex, higher strength microstructures. For
example, as indicated above, the “grain size” strengthening in the IAC steels will almost always be
overestimated. Furthermore, any discrepancy between the strength observed in IAC microstructures and
that predicted by the equations developed for polygonal ferrite, Equation 2, but applied to IAC structures
is often immediately attributed to precipitation hardening. In most cases, the precipitation hardening
increment is calculated by assuming that the difference between the measured and calculated strength,
based on grain size and solutes, is caused by particle hardening. However, this claim of a given level of
precipitation hardening is almost never supported with independent verification, e.g., through the
combination of the application of the Orowan-Ashby precipitation hardening relationship[9,14] using
inputs from solubility relationships and especially thin foil TEM.

The goal of this present paper is to describe what is considered a new and better method of identifying
and characterizing the individual ferrite components in complex microstructures of the kind observed in
high strength hot rolled strip or plate processed steels subjected to IAC with low WET. This approach
can also be used in: (a) the analysis of multi-phase microstructures as found in the advanced high
strength sheet steels (AHSS) with dual-phase, TRIP and Complex Phase microstructures; (b) annealing
studies where the IQ value is related to the level of recovery or recrystallization of individual grains, and
(c) in heat affected zones of welds.

New Metallographic Technique based on EBSD Image Quality

Modern MA steels subjected to IAC with low WET can exhibit a final microstructure containing at least
five distinct components: four types of ferrite: (i) polygonal ferrite, (ii) non-polygonal ferrite, (iii)
acicular ferrite, (iv) bainitic ferrite, and (v) the complex microconstituent martensite-austenite-cementite
(MAC). Although the types of ferrite differ from one another in many ways, the one systematic
difference is in their own particular dislocation density, which increases with falling transformation
temperature. Hence, the intrinsic dislocation density would be expected to increase as we go from
polygonal to non-polygonal to acicular to bainitic to martensitic. An example in going from polygonal to
acicular ferrite is shown in Figure 3.[15,16] Likewise, we would expect the line broadening of the
Kikuchi peaks to spread and peak heights to fall in a similar fashion; polygonal ferrite not much, non-
polygonal a little more, acicular more and bainite more still.

1011 250
Equiaxed grains
Plate-like grains
Mean
Pa
174 M
cm-2 S,
sity, Y
Den Δ
ion ed
1010
ocat 78 at
Disl m
ti
Es
55

109 25
650 700 750 800
Transformation Temperature, °C
Figure 3: Dislocation densities as a function of transformation temperature. Fe-0.21Ti-0.08C.[16]
In a recent series of studies, a technique based on EBSD image quality (IQ), as defined by the elastic
distortion of the diffracted intensity associated with the Kikuchi bands, has been used to identify and
quantify the different types of ferrite, found in complex microstructures, and varying in intrinsic
dislocation density.[17] In this technique, IQ values are determined from each grain in the analyzed area
of the sample and IQ values from the grain boundary regions are eliminated. The resulting, corrected
values are then normalized such that all IQ values fall between zero and 100. This total histogram of
corrected and normalized values is then de-convoluted into multiple peaks, where the number of peaks
represents the number of types of ferrite contributing to the total. This is called the Multi-peak analysis
or model. In the final analysis, the area under a given peak represents the amount of that particular
phase and the location of the peak on the IQ axis (abscissa) gives the peak IQ value or an index of the
dislocation density for that particular phase or microconstituent.

Application of the New IQ Technique

As-Coiled Microstructure of High Strength Low Alloy (HSLA) Steel Hot Band

In a recent study of the origins of variability of strength in the hot strip mill processing of 0.08C-1.4Mn-
0.038Nb-0.050V HSLA steel with a yield strength near 490MPa, it was found the final microstructure
was strongly influenced by the finishing and coiling temperatures. An example of these results is shown
in Figure 4,[18] which shows the influence of finishing and coiling temperatures on the final
microstructure, as qualitatively described by optical microscopy. One such microstructure was shown
above in Figure 2 for low temperature finish rolling and low coiling temperatures, and it is clear that the
final microstructure is a mixture of several different types of ferrite. The corresponding EBSD analysis
of the specimen used for Figure 2 are presented in Figures 5 and 6, where Figure 5 shows the
distribution of grain boundary misorientations and IQ in terms of gray scale, and Figure 6 shows the IQ
analysis using the multi-peak model. For a complex microstructure like this, the visual identification of
the ferrite types based on the morphological descriptions might be possible, but any quantitative
measurement is extremely difficult, if not impossible. It is therefore nearly impossible to get an accurate
measure of the volume fractions of the various types of ferrite present by using any of the traditional
microstructural analyses in complex microstructures.
700

FP+P FP+ FN +P FP+P FP+P FP+P

650
FN+FB+P

600
C oiling Tem p (C )

FN+FB FN+FB+C FN+FB FN+FB+C

o

550

High Reheating, High Roughing High Reheating, Low Roughing

500
700

FP+P FN+FB+P FP+P FP+P FP+P

650

600

FN+FB+P FN+FB+(P) FN+FB+C FP+P FN+FB + C

550

Low Reheating, High Roughing Low Reheating, Low Roughing

500
900 950 1000 900 950 1000
o
Finishing Temp ( C)

Figure 4: Mapping of the phase mixtures for different reheat and

roughing temperatures in CT- FT space. Hot-rolled Nb+V steel.[18]
EBSD maps of the microstructure are shown in Figure 5, in which (a) is the grain boundary and subgrain
boundary misorientation map and (b) is the gray scale image quality map. It is observed that the grains
with interior subgrain boundaries always have relatively lower IQ values compared to those without
subgrains. The ferrite types formed at low temperatures, e.g. non-polygonal, acicular and bainitic, have a
higher degree of lattice distortion and a larger density of subgrain boundaries. Even for the grains
without subgrain boundaries, the IQ values vary noticeably from one grain to another. This variation of
IQ values implies a difference in lattice imperfection, e.g., dislocation density, which is very much
determined by the mechanism and temperature of formation of the ferrite type. Consequently, the
differentiation of IQ values caused by the lattice imperfection serves to identify the different types of
ferrite.

Figure 5: EBSD maps of the HSLA steel hot band microstructure shown in Figure 2.
(a) The grain boundary and subgrain boundary misorientation map, where the thick curves are for grain
boundaries with misorientations larger than 15 and the thin curves are for those from 2 to 15,
(b) The gray scale image quality map of the identical area of (a).

The result of applying the multi-peak model to the microstructure presented in Figure 2 is shown in
Figure Figure 6. A total of five normal distribution peaks is summed up and the difference from the
experimental data curve is around 1.7 pct. Because the traditional classification of ferrite is mainly based
on the visualized grain morphology and the IQ method relies on the lattice imperfection, the
correspondence of the multi-peaks in Figure 6 to the particular ferrite types is based on two assumptions.
The first assumption is that each component of the final microstructure has its own, individual
characteristic peak in the IQ distribution curve. The second is that the mean value for each IQ peak
decreases with falling transformation temperature or increasing lattice defect density of each type of
ferrite. Thus, some correspondence of the microstructural variation and the IQ distribution curve may be
possible, as indicated in Figure 6.

The IQ analysis itself is a method of characterizing microstructures based on a 3-D view of lattice
imperfection, as opposed to the traditional analysis based on 2-D surface visualization.
 0.18
Expe rime ntal Data with GBR Filte re d
0.16 Po lyg o nal Fe rrite
No n-Po lygo nal Fe rrite
0.14 Ac ic ular Fe rrite
Bainitic Fe rrite
0.12 Carbon Ric h Mic roc ons titue nts
S um of all Pe aks
0.1

Fraction
0.08

0.06
40.6%
0.04 3.4%
8.9% 36.3%
0.02
10.8%
0
0 10 20 30 40 50 60 70 80 90 100 110
No rmalize d Imag e Quality

Figure 6: The IQ analysis of the HSLA hot band microstructure shown in Figure 2
using the Multi-Peak model.

Microstructure of a Dual-Phase Advanced High Strength Steel After CGL Processing

In this study, two low carbon dual-phase steels were hot rolled, cold rolled and intercritically annealed
and further processed using a CGL simulation.[18,19] The steels had a composition of 0.06wt%C,
1.5%Mn, 0.45%(Cr+Mo), and either 0.02 or 0.04%Nb, depending on whether 590 or 780 properties
were expected. After cold rolling, the steels were heated to the intercritical annealing temperature,
soaked and then cooled at different cooling rates to 460°C, held for various times and then quenched to
RT. The final microstructure consisted of five contributions: (i) recrystallized ferrite formed from the
cold rolled ferrite in the cold band, (ii) unrecrystallized ferrite from the cold rolled ferrite in the cold
band, (iii) new ferrite that formed during cooling from the intercritically formed austenite, (iv)
martensite that formed from the intercritically formed austenite, and (v) some undissolved NbCN
inherited from hot band. Examples of these microstructures taken with SEM are shown in Figure 7.[19]

(a) (b) (c)

Figure 7: Typical SEM micrographs observed in both steels after the CGL simulation treatment.
(a) BDP-580 3°C/s; (b) BDP-790 8°C/s (c) BDP-790 15°C/s.

The IQ-Multi-peak analysis of these structures revealed three peaks, for recrystallized ferrite, for
unrecrystallized ferrite and for martensite. Note that the unrecrystallized peak contains both the non-
recrystallized cold rolled ferrite plus the new ferrite formed upon cooling from the annealing to the zinc
pot or overaging temperature. The 590 DP steel microstructure is shown in Figure 8. This work has
shown that the 590 DP steel exhibited about 19% martensite, Figure 8, while the 780 grade had about
30%, Figure 9. These quantities are remarkably similar to those found in an earlier study due to Bucher
and Hamburg, where they found approximately 15 and 35%, respectively, were required for these
strength levels.[20] Of course, the IQ analysis would not appear to be critical in the study of DP steels,
since the amount of martensite is not only critically important to properties, but also is easily observed
and measured by typical, quantitative optical microscopy. However, there is no other way to easily
distinguish among the different kinds of ferrite that exist entering the zinc pot; i.e., cold rolled ferrite
that has recrystallized in the anneal, cold rolled ferrite that has not recrystallized in the anneal, and new
ferrite that forms at low temperature during the cooling portion of the process. The first ferrite would
have a high IQ, while the other two would have a low one.

Figures 8 and 9: IQ analysis of the dual-phase microstructure after CGL processing.

Microstructure of a TRIP Advanced High Strength Steel After CGL Processing

An investigation has recently been completed in which a low carbon, low Si, high Al, Nb-bearing TRIP
steel was processed using an intercritical anneal followed by a CGL processing simulation.[19] The
steels investigated contained 0.15wt%C-1.5%Mn-0.3%Mo-0.03%Nb and Al varying from 0.05 to 1.0%.
After hot rolling, coiling at 550°C and cold rolling, the steels were heated to the soak temperature, held
for one minute at temperatures ranging from 750 to 860°C, then cooled at different rates to 450°C, held
for various times then cooled to RT, again at different rates. The microstructures were observed with
OM, SEM, TEM and the new EBSD-IQ technique. The retained austenite values were checked by
magnetometry. The conditions examined were: (i) as-intercritically annealed and WQRT; (ii) annealed
and cooled to 450°C and WQRT; and (iii) annealed and cooled to 460°C, held for various times at
450°C and cooled to RT.

An example of the application of the EBSD-IQ technique incorporating the Multi-peak Model is shown
in Figure 10, which compares the makeup of the microstructure in the 0.05Al and 1.0Al steels after a
one minute anneal and cooling at 15°C/sec to 450°C, followed immediately by water quenching to RT.
The influence of the Al content on the A3 line and kinetics of annealing are striking. The Al appears to
shift the A3 line to the right and upward, leading to less austenite but probably austenite of a higher
carbon content. The high Al content also seems to accelerate the annealing kinetics, with much more
recrystallized ferrite in the higher Al steel.
 EBSD (IMAGE QUALITY) ANALYSIS

ALLOY 0.05Al-550 ALLOY 1%Al-550

TIA=750 C, 60s, 15C/sec to 450C, Quench TIA=770 C, 60s, 15C/sec to 450C, Quench
0.25 0.18
EBS D Data witho ut GB Contribution EBSD Data without GB Contribution
RXD Ferrite Fe rrite RXDl Fe rrite
0.16
Non-RXD Ferrite No n-RXD Fe rrite
0.2 Martens ite
0.14 Martens ite
Sum o f all s imulated co ntributions
Re tained Aus tenite e
0.12 Sum of all s imulate d c ontributions
37.17

Po pulatio n
0.15 50.97%
0.1

Frequency
26.55 36.28
0.08
29.13%
0.1
0.06

0.04
0.05 3.52 % 16.38%
0.02

0
0 0 10 20 30 40 50 60 70 80 90 100 110
0 10 20 30 40 50 60 70 80 90 100 110 IQ
IQ

Austenite (Martensite) = 26.55% Austenite (Martensite) = 16.38%

Recrystallized Ferrite= 37.17% Recrystallized Ferrite= 50.97%
Non-RXD + New Ferrite= 36.28% Non-RXD + New Ferrite= 29.13%
New Ferrite EBSD (IMAGE
= 36.28-27.78= 8.5% QUALITY) ANALYSIS
New Ferrite = 29.13-16.9= 12.23%
Retained Austenite= 0% Retained Austenite=3.52%

Figure 10: IQ multi-peak quantitative analysis of TRIP steels varying Al content.

ALLOY 1%Al - CT=550

TIA=770 C, 60s, Quench TIA=860 C, 60s, Quench

0.2 0.25
EBS D Data without GB Co ntributio n
EBSD Data without GB Contribution
RXD Ferrite
0.18 RXD Ferrite No n-RXD Fe rrite

0.16 No n-RXD Fe rrite 0.2

Martens ite
S um o f all s imulated contributions
Martens ite
0.14
Sum of all s imulate d c ontributions
0.12 57.5% 32.4%
Po pulation

0.15
Frequency

0.1
52.8%
0.08
0.1

0.06 16.9%
10.1%
0.04
30.2% 0.05
0.02

0
0 10 20 30 40 50 60 70 80 90 100 110 0
0 10 20 30 40 50 60 70 80 90 100 110
IQ
IQ

Austenite (Martensite) = 30.2% Austenite (Martensite) = 57.5%

Recrystallized Ferrite= 52.8% Recrystallized Ferrite= 32.4%
Non-recrystallized Ferrite= 16.9% Non-recrystallized Ferrite= 10.1%
Normalized RXD Ferrite = 75.7% Normalized RXD Ferrite = 76.2%
Figure 11: IQ multi-peak quantitative analysis of 1.0Al TRIP steel.

The influence of intercritical annealing temperature on the microstructure found after cooling to 450°C
at 15°C/sec and WQRT with no holding time at 450°C is shown in Figure 11 for the steel containing
1.0Al. As expected, the higher annealing temperature resulted in a lower carbon content in the
intercrically formed austenite and subsequently lower hardenability. The amount of new ferrite formed
upon cooling was about 44% with the 860°C anneal, but only 12% for 770°C. Annealing at 860°C
would also be expected to lead to more austenite being formed. Hence, the sum of the new ferrite plus
the martensite was 57% for the 860 anneal but only about 22% for the 770°C anneal, even taking the
retained austenite into account. The formation of the austenite appeared to occur at the expense of some
of the recrystallized ferrite in the specimens annealed at 860°C. Normally, it would be expected that
higher annealing temperatures would result in more recrystallized ferrite, not less. However, since the
driving force for austenite formation is much higher than for the recrystallization of ferrite, the
formation of the austenite would take preference and would be especially noticeable in the specimen
annealed at 860°C. Also, nearly 4% retained austenite was found at 450°C, in the absence of isothermal
holding at 450°C, in the samples annealed at 770°C. This is no doubt related to the initially higher
carbon content of the initial austenite formed at 770°C.
0.14
EBSD Data without GB Contribution
RXD Ferrite
0.12 Non-RXD Ferrite
Bainite
Retained Austenite
0.1 martensite
Sum of all simulated contributions
48.31%
0.08

ion
ulat 0.06
Pop
31.33%
0.04
6.83%
0.02 9.86%

0
0 10 20 30 40 50 60 70 80 90 100 110
IQ
Figure 12: IQ multi-peak quantitative analysis of 1.0Al TRIP steel, TIA=770°C (60s), 15°C/s to 450°C,
120s-AC. Recrystallized Ferrite= 48.31%, Non-RXD + New Ferrite= 31.33%, Bainite = 6.83%,
Retained Austenite=9.86%, Martensite =3.67%

The final microstructure found after annealing at 770C, cooling to 450°C at 15°C/sec and air cooling to
RT after a holding time of 120 sec at 450°C is shown in Figure 12 for the steel containing 1.0Al. The
microstructure, revealed both by IQ-Multi-peak and supported by standard metallography, showed about
48% recrystallized ferrite, 31% unrecrystallized ferrite (sum of old recovered and new ferrite), about 7%
bainite, 10% retained austenite and 3.7% martensite. As indicated above, each of these features can be
observed and identified using standard techniques, but obtaining a quantitative estimate of the amounts
of each is not possible in any other reliable way.

Conclusions

From the time of Dube in the 1950s[20,21] to the present, researchers have known that ferrite and other
microconstituents can assume many morphologies, depending on the composition and processing of
steel. The various types of ferrite found in low carbon steel have been described in numerous papers,
often with supporting illustrations and other documentation. It has been common for attempts to be
made to quantify complex microstructures, using point counting and other standard quantifying
techniques. However, these attempts are not satisfying because they are predicated on being able to
distinguish among the different types of ferrite, which is, in fact, not possible based on conventional
metallography. In this paper, a new approach has been presented, the IQ-Multi-peak model, which
allows the components of complex microstructures to be quantified. This technique, in its several forms,
reveals much information regarding the microstructure, itself, plus valuable information regarding the
level and distribution of crystalline defects in the microstructure. Its application to real problems in
steels has begun, and its eventual use in analyzing more complex microstructures such as found in
higher strength steels, welds, hot and cold deformation and forming operations is under development.

Acknowledgements

The authors would like to thank the Basic Metals Processing Research Institute, Department of
Materials Science and Engineering, University of Pittsburgh and its industrial and governmental
sponsors for the generous support of this work.

References

1. HSLA Steels, Technology and Applications, Conference Proceedings of International Conference on

Technology and Applications of HSLA Steels, Philadelphia, PA, USA, ASM, Metals Park, OH, USA
(1984).

2. A.J. DeArdo, "The Influence of Thermomechanical Processing and Accelerated Cooling on Ferrite
Grain Refinement in Microalloyed Steels", Accelerated Cooling of Steel, (Warrendale, PA, TMS-AIME:
1986), p. 97.

3. R. Colas and C. M. Sellas, "Computed Temperature Profiles of Hot rolled Plate and Strip During
Accelerated Cooling", Proceedings of the International Symposium on Accelerated Cooling of Rolled
Steel, Winnipeg, Canada, 1987, Vol. 3, Pergamon Press, London, pp 121-130

4. D. Ellerbrock, et. al., "Characterization of wide-angle spray nozzles for use in accelerated cooling of
hot steel bodies", ibid, 147-157

5. A. J. DeArdo, “Accelerated Cooling: A Physical Metallurgy Perspective,” Canadian Metallurgical

Quarterly, vol. 27, no.2, (1988), 141-154.

6. H.G.Hillenbrand, et al, “Development of Large-Diameter Pipe in Grade X-100,” Pipeline Technology,

Vol.1, 2000, 483-496.

7. E. O. Hall, “The deformation and ageing of mild steel: III. Discussion of Results,” Proc. Phys. Soc. B,
64 (1951), 747-53.

8. N. J. Petch, “The cleavage strength of polycrystals,” J. Iron Steel Inst., 174 (1953) 25-8.

9. T. Gladman: in Proc. Microalloying '75, New York, NY. 1977, Union Carbide Corporation, 32-54.

10. W.C. Leslie: The Physical Metallurgy of Steels, 343; 1981, New York, NY, Hemisphere Publishing.

11. K.J. Irvine, F.B. Pickering and T. Gladman: J. Iron Steel Inst.,1967, 205, 161-182.

12. W. M. Hof, et al., New High-Strength Large-Diameter Pipe Steels, Proc Intl Conf, HSLA Steels ’85
(Beijing), (ASM, Materials Park, OH, USA, 1986), 467-474.

13. J. Wu, PhD Thesis, University of Pittsburgh: Pittsburgh, PA, USA, 2005.

14. M.F. Ashby: Phil. Mag., 1970, 8(21), 399-424

15. V. Thillou, et al., “The Strength Properties of Hot Strip HSLA Low Carbon Steel,” Proc. 41st
MWSP (Baltimore), Vol. XXXVII, (ISS, Warrendale, PA, USA, 1999), 471-480.

16. W.F. Smith: ‘Structure and Properties of Engineering Alloys’, 1981, New York, McGraw-Hill.

17. Jinghui Wu, et al., “An Advanced Understanding of the Polygonal Ferrite Microstructures of IF
Steel Using Image Quality Analysis,” Processing and Fabrication of Advanced Materials XIV With
Frontiers in Materials Science 2005: Innovative Materials & Manufacturing Techniques, Proc. Materials
Science & Technology 2005 Conference, (ASM, AcerS, AIST, AWS, and TMS, 2005), 283-306.

18. Jose E. Garcia-Gonzalez, MS Thesis, University of Pittsburgh, PA, USA, 2002.

19. C. I. Garcia, K. Cho, Y. Gong, T. R. Chen and A. J. DeArdo, “Development of High Strength, Low-
Carbon, Nb-Bearing Dual-Phase Steels For Production on Continuous Galvanizing Lines,”
Developments in Sheet Products for Automotive Applications, Organized by James R. Fekete and Roger
Pradhan, Materials Science & Technology 2005, (ASM, ACerS, AIST, AWS and TMS, 2005), 77-86.

20. J. H. Bucher and E. G. Hamburg, SAE Trans 86.730, Sect. 1, 1977.

21. C. A. Dube, PhD Thesis, Metallurgical Engineering, Carnegie Institute of Technology, Pittsburgh,
PA, 1948.

22. R. F. Mehl and C. A. Dube, Phase Transformations in Solids, Wiley, New York, 1951, p545.