Barekar2014P Overview of Twin Roll Casting of Al Alloys
Barekar2014P Overview of Twin Roll Casting of Al Alloys
Barekar2014P Overview of Twin Roll Casting of Al Alloys
To cite this article: N. S. Barekar & B. K. Dhindaw (2014) Twin-Roll Casting of Aluminum Alloys – An Overview, Materials and
Manufacturing Processes, 29:6, 651-661, DOI: 10.1080/10426914.2014.912307
Taylor & Francis makes every effort to ensure the accuracy of all the information (the “Content”) contained
in the publications on our platform. However, Taylor & Francis, our agents, and our licensors make no
representations or warranties whatsoever as to the accuracy, completeness, or suitability for any purpose of the
Content. Any opinions and views expressed in this publication are the opinions and views of the authors, and
are not the views of or endorsed by Taylor & Francis. The accuracy of the Content should not be relied upon and
should be independently verified with primary sources of information. Taylor and Francis shall not be liable for
any losses, actions, claims, proceedings, demands, costs, expenses, damages, and other liabilities whatsoever
or howsoever caused arising directly or indirectly in connection with, in relation to or arising out of the use of
the Content.
This article may be used for research, teaching, and private study purposes. Any substantial or systematic
reproduction, redistribution, reselling, loan, sub-licensing, systematic supply, or distribution in any
form to anyone is expressly forbidden. Terms & Conditions of access and use can be found at http://
www.tandfonline.com/page/terms-and-conditions
Materials and Manufacturing Processes, 29: 651–661, 2014
Copyright # Taylor & Francis Group, LLC
ISSN: 1042-6914 print=1532-2475 online
DOI: 10.1080/10426914.2014.912307
Since its invention by Sir Henry Bessemer in 1865, twin-roll casting (TRC) has been the subject of extensive research, not only to develop
the technology but also to achieve an understanding of microstructural evolution. The present review confines itself to the literature on
process aspect, modeling, and quality issues. Initially, the principles of the process are outlined. Modeling of fluid flow, heat transfer,
and microstructural evolution, surface and internal defects in TRC of aluminum alloys are next discussed. The role of process parameters
on solidification during casting is reviewed. The controls of grain structure by melt treatment are also discussed in brief.
INTRODUCTION for few wide freezing range alloys [7]. There are also
One of the major consumers of wrought aluminum issues of performance of products made from TRC
products is automotive industry [1, 2], where weight strips due to centerline segregation [8].
reduction is a primary goal. Aluminum with only one-third These issues have been addressed by considerable
of the density of steel and better strength to density ratio research and development to further understand and
could provide weight reduction of nearly 50% when used improve the process. Thus this review has been undertaken
in place of steel. The major barrier to the widespread use to evaluate the past work and state of art of TRC process
of aluminum sheets in high volume is its high cost of pro- in terms of technology and the fundamental aspects.
duction. Production of aluminum sheets by continuous
casting route rather than the conventional direct chill STRIP CASTING
(DC) casting and hot mill method offers an opportunity There are two types of strip casting processes used in the
to substantially reduce the cost [3]. Because of the econom- production of sheet for Al alloys. The oldest and classical
ics and metallurgical advantages offered by twin-roll cast- form is twin belt casting (TBC). This is suitable for thicker
ing (TRC) process, it has become widely popular in slab gauges varying from 10 to 25 mm [3, 9]. Here, the
aluminum industries [4]. Currently TRC process offers dis- metal solidifies between two water-cooled counterrotating
tinct advantage in lowering of greenhouse emissions. The belts. The process has been successfully used for wide
capital investment for TRC is significantly lower than the freezing range Al alloys. The cooling rate in TBC during
conventional DC casting and hot rolling process. TRC solidification is 102 K=s. There is no significant defor-
requires low energy consumption, limited space require- mation of the strip during solidification process. Therefore
ment, and it also offers possibilities of easy diversification. defects associated with deformation during semisolid stage
In general metallurgical terms, TRC is expected to are not encountered. However, the reduction in thickness
have refined microstructure, fine intermetallic particles, from casting to final product may be as high as 60% [3].
and increased solid solubility [5], which is advantageous
for mechanical properties. Despite the above-mentioned
advantages, the aluminum industry is still looking for Principles of TRC Process
improving the economic advantage in terms of pro- The second set of process is TRC which is the topic of
ductivity to remain competitive with other processes this review. The schematic of horizontal twin-roll caster
and emerging flat roll products [6]. The quality has been is shown in Fig. 1, which is derived from Ferry [5]. It is
assessed for limited alloy compositions but not universal pertinent to note that industrial TRC of Aluminum is
across the wide spectrum of compositions of Al alloys. focused mainly on horizontal or near horizontal tech-
Like, for alloys with narrow solidification range, TRC nology. In TRC, the thickness of the processed sheet is
has been successful, however processing is still restricted generally between 3 and 7 mm, although less than
3 mm thickness is possible at higher casting speed [10,
Received December 13, 2013; Accepted March 24, 2014
11]. For casting high-quality thin gauge strips, TRC
Address correspondence to B. K. Dhindaw, Indian Institute of requires a set of robust rolls capable of withstanding
Technology Kharagpur, Kharagpur, India 721302; E-mail: dhindaw@ the high roll torques and roll separation forces. It has
gmail.com been generally predicted that operating point of a
Color versions of one or more of the figures in the article can be found twin-roll caster is determined by the force exerted by
online at www.tandfonline.com/lmmp. the rolls on the strip (expressed as tons=m of strip
651
652 N. S. BAREKAR AND B. K. DHINDAW
and different combinations of microstructure across the geometry that in turn depends upon the alloy type, roll
strip thickness were obtained varying from lamellar eutectic material, roll diameter, rolling speed, and pouring tem-
to wavy=flowery morphologies. It was concluded that TRC perature. The three important parameters that decide
could be utilized for direct production of layered composite the solidification structure are related to setback distance,
strips provided the casting speed is sufficiently high. sump depth, and mushy zone depth. The three lengths are
Although it is possible to cast composites by TRC, shown in Fig. 2. Solidification commences by nucleation
limited attempts have been made due to complicated at the roll surface as the melt exits the tip of tundish
fluid flow and heat transfer during solidification in TRC. and proceeds by dendritic growth. The distance required
to solidify the strip as it exits the nozzle is the sump depth,
and the distance between the tip of nozzle and the roll nip
MOLTEN METAL SUPPLY TO THE TWIN-ROLL CASTER is termed as setback. The degree of deformation during
It was recognized long back that proper design of TRC is less for bigger diameter rolls at the same setback
tundish is important to maintain precise control of mol- and roll gap whereas it is more for smaller roll gap for the
ten metal head in the tundish and to ensure a regular, same setback and roll diameter [5].
uniform, turbulent-free melt to the roll bite at any cast-
ing speed. A typical metal feeding system to the roll gap
for TRC of aluminum alloys is given in reference [29]. MODELING OF FLUID FLOW AND HEAT
Inclusions in the melt can be reduced by passing the TRANSFER DURING TRC
molten metal through degassing and filtration units No clear perspectives about the exact geometrical
[12]. In TRC of aluminum alloys, temperature of the
Downloaded by [New York University] at 16:01 06 January 2015
AA3003 alloy were used to do the calculations. No rich material that form on the strip surface and also
details about the validation of the model with actual contain higher concentration of intermetallic particles.
casting situation were given although it was mentioned Surface bleed example is shown in Fig. 5. Not all surface
that it helped in improving the process. In a communi- bleed had similar structure. It was suggested [42] that
cation, Das et al. [40] show the temperature distribution initially a thin film forms as the molten metal solidifies
in the strip during TRC by numerical modeling. It was on the roll. As it approaches the roll gap, it meets
suggested that the temperature at the center of the speci- material that is traveling more slowly due to backward
men is higher compared to surface for higher rolling slip, and the resultant differences in velocities cause the
speed of 5 rpm and was attributed to lower contact time. film to lose contact with the roll surface. Because of
A comprehensive computational fluid dynamics the load acting, semisolid material in this region forces
mathematical model for high-speed TRC of molten the solute rich liquid to the gap formed on the surface.
Al-33 wt.% Cu alloy was developed using FLUENT The formation of surface bleed mostly depends on the
6.3. 16 platform [41]. The model was validated exper- freezing range of the alloy. In summary, surface bleeds
imentally using Jackson-Hunt plot for Al-33 wt.% Cu, have been identified as the major defects in twin-roll cast
and the effect of process parameters on solidification aluminum sheets. It has been reported [45] that high
front speed was studied. The details of the modeling casting speed (low casting load), thin strip thickness,
procedure and experiment conducted for validation are alloys with freezing ranges 10–30 C and reasonable
given in reference [38]. Interesting velocity profile of the amount of liquid present prior to completion of solidifi-
movement of liquid melt in the molten pool was pre- cation promote the formation surface bleeds.
dicted. Figure 3 shows the schematic of velocity profile Forbord et al. [46] studied the formation of surface
that correlates with general flow patterns expected in segregates during TRC of AA5052 aluminum alloys by
molten metal pool in TRC. The schematic is derived employing a technique termed as stop sample. In this,
from reference [41]. The reversal of the flow direction
in the middle of the pool before it entered the roll bite
region was shown. The solidification front speed and
solidified shell thickness for roll speed varying from 3
to 500 rpm were presented. The experimental validation
essentially involved calculations of interlaminar spacing
by using Jackson-Hunt relationship for this alloy. As
predicted by the model, the microstructure of the solidi-
fied Al-33 wt.% Cu strip was not uniform, and two dis-
tinct zones were observed. This was rationalized in
terms of variation of solidification speeds. However
above work is related with vertical TRC.
rolls. The strip stuck in the roll gap region was charac- load would be function of rolling speed. They went on to
terized for microstructure development to investigate draw defect limit diagrams. The tendency of segregation
segregation. The studies were supplemented in a limited increases with increasing casting speed and casting gauge
way by modeling. It was proposed that when a certain [47]. At given casting conditions, the centerline segre-
productivity=casting speed is exceeded, the enriched gation in twin-roll cast strip increases as the solute con-
liquid in interdendritic regions connects and eventually tent in the alloy (and hence freezing range) increases.
becomes transport paths to the surface. This flow is For the central region, the researcher [44] believed that
further enhanced by low-pressure surface zones. These enriched liquid in the deep sump transform to equiaxed
segregations appear as protrusions on the surface. How- segregates. However there is no direct evidence or con-
ever, these protrusions bend towards casting direction sistency in this mechanism. They generalized the proces-
and are eventually squeezed into the underlying surface. sing factors like if casting gauge and applied load are
Working with AA3105 aluminum alloy, Ch. Cras et al. small, bleeds, and center segregates form. The centerline
[44] attributed formation of surface defects to the forma- segregates are formed at lower temperature than the sur-
tion of buckle formed due to local velocity differences face bleeds and contain less primary phase. They con-
induced by backward motion of the strip when roll cluded that for a fixed alloy composition and casting
pressure is applied. The interdendritic liquid is squeezed gauge, high-speed casting definitely promotes the pro-
towards the surface into free space created by the duction of micro defects.
buckle. This liquid fills up the empty space to form dro- Kim et al. [48] demonstrated that using sufficient roll
plets and instantaneously solidify when in contact separating force can lower the occurrence of defects in
with the roll. Later the passage in the roll bite deforms the central region even for wide freezing temperature
the blister and gives the bleed saucer like shape as seen range AA6022 alloy cast during high-speed TRC. How-
the Fig. 5. ever for alloys having large content of intermetallics like
The separation of the shell from the rolls during TRC those containing Si and Mg, low level of force between
of aluminum alloys can increase the surface temperature rolls gave better structure with less harmful distribution
of the strip due to reduced heat transfer rate. This of intermetallics [49]. For a twin-roll cast Al alloy con-
results in formation of large intermetallic and coarse taining 1.04% Si and 0.56% Mg, Das et al. [50] have
grain structure as well as oxide on the rolls at the point reported the centerline segregation of Si, Mg, and Fe
of separation. During extended operation, the oxide at different speeds (3, 4, and 5 rpm), although no quan-
areas on the roll surfaces tend to develop transverse titative segregation data is provided. The sheet cast at
bands of coarse structure (i.e., ripple effect) on the high roll speed also exhibited rough surface.
strip surface [5]. On the other hand, Birol [51] has focused on the issues
of macrosegregation in twin-roll cast strips in a different
Internal defects. The centerline segregation in the way. In an interesting study, differential scanning calor-
TRC strip is due to the segregation of solute elements imetry (DSC) was employed to analyze the solidification
which are pushed to the center from the two opposite behavior of twin-roll cast AA1050, AA3003, AA5754,
advancing solid=liquid interfaces of the growing grains. and AA6016 aluminum alloys. Alloys AA3003 and
Under the roll pressure, the solute-rich cold liquid AA5754 having widest solidification range as deter-
squeezes out from the central region to the hot liquid mined by DSC showed maximum macrosegregation in
region and pumps back the hot liquid to the mushy industrially cast strips by TRC. The center line segre-
region where it melts the primary solid. When liquid gates were hardly altered by annealing [52]. Therefore
656 N. S. BAREKAR AND B. K. DHINDAW
Downloaded by [New York University] at 16:01 06 January 2015
FIGURE 6.—Internal defects (indicated by arrows) observed in twin-roll cast Al alloys (a) channel segregates (b) deformation segregates and (c) banded
structure.
to handle these alloys for TRC would require very change in cooling rate across the section. The intermetal-
judicious control of casting parameters such as setback, lics also doubled in size from surface to center. An array
tip orifice, casting gauge, and the casting speed in order of segregates were also seen spaced more or less equally
to minimize the defects. from one edge to the other and roughly marked the cen-
terline of strip. The rolling component of TRC is defi-
Buckling. Buckling in thin twin-roll cast sheets nitely responsible to these defects. Beng et al. [53]
(thickness < 3 mm) has been reported by Lockyer et al. showed that microstructure was function of the strip
[42], where one side of the sheet is tight and the other thickness for twin-roll cast AA5052 alloys. For thicker
is long. For the wide freezing range alloys, the long edge strips (5 mm), dual grain structure consisting of coarse
deformed heavily compared with the short edge. They dendritic head and a cellular tail was seen. Cast strips
related the buckling to the nonuniformity of the solidifi- of 1.9 mm thick showed finer and purely equiaxed struc-
cation process resulting in more forward slip over the ture. The equiaxed grains were somewhat elongated. A
part of the strip. The buckling was found to be alloy spe- deeper mushy zone with a more vertical solidification
cific and could be reduced or eliminated by increasing front can explain these results.
the coiler tension. Okayasu et al. [54] while studying mechanical proper-
ties of TRC sheets of Al–Si–Cu alloys found that micro-
structural characteristics of the TRC samples were
Microstructural Evolution dominated by an anisotropic microstructure formed by
As cast microstructure of TRC strips. In a separate a fine dendritic columnar and equiaxed grains in longi-
study [52] of a twin-roll cast AA8011 alloy (Fe 0.7, tudinal directions. They associated such microstructural
Si 0.77, Cu 0.038, Mn 0.006) widely used to pre- characteristics with the casting conditions like metal
pare thin foils, some interesting observations regarding temperature, mold temperature, and casting speed.
the microstructure variation across the thickness of the Slamova et al. [55] studied the microstructural character-
strip were seen. It was observed that secondary dendrite istic of AA5052 and AA5182 alloy prepared by DC cast-
arm spacings (SDASs) were finer along the outer surface ing and TRC processes. It was also reported that TRC
and gradually coarsened towards interior indicating process gave finer microstructure for both the alloys
TWIN-ROLL CASTING OF ALUMINUM ALLOYS 657
compared to DC casting. The second phase particles for the alloys with large solidification range usually dif-
were large in number in TRC and well dispersed. It ficult to twin-roll cast. They could cast strips thinner
was found that AA5052 showed more anisotropic struc- than 3 mm and at speed of 60–180 m=min. However, sev-
ture than AA5182 which had large content of Mg and eral surface defects were seen. The defects could be mini-
Mn. Although the microstructure was fine, TRC showed mized after heat treatment. The structure in the middle
wide variation in grain size across the section of the of strip was globular. The fineness of the microstructure
strip. The variation in microstructure and texture were was attributed to low superheat and cooling ability of
analyzed for solid solubility of elements due to different roll. While Haga et al. [14, 17, 19, 26] have been quite
solidification rate. successful with innovative technologies to process strips
Das et al. [40] have reported that SDAS is large for of alloys with wide solidification range or composites
higher rpm than lower rpm twin-roll cast alloys. While under different conditions by TRC, the microstructural
the alloys twin-roll cast at roll speed of 3 and 4 rpm issues like macrosegregation and microstructural
showed elongated grains and the one cast at 5 rpm inhomogeneities have not been addressed.
showed equiaxed grains. They attributed the equiaxed
grain formation at 5 rpm to dynamic recrystallization CONTROL OF GRAIN STRUCTURE IN
taking place due to higher heat retention [50]. The TRC BY MELT TREATMENT
microstructure was also finer at this roll speed. The Grain Refiner Addition
results and discussion does show interesting pattern in
the microstructure development with increasing roll The control of grain structure by the deliberate
speed. However, the mechanism of microstructural addition of grain refiner for heterogeneous nucleation
Downloaded by [New York University] at 16:01 06 January 2015
evolution has some contradictions. into the melt is a well-established practice. Another
expectation from grain refiner addition is the reduction
Microstructure of semisolid TRC strips. Haga et al. of macrosegregation which is a major issue in TRC.
[14] in a novel attempt used unequal diameter twin-roll Commercially available Al-3Ti-0.15C grain refiner was
caster where upper roll was four times smaller diameter employed in the TRC of AA8111 foil stock [56]. The
than lower roll. They cast the strips by introducing the grain refining performance was compared with that of
alloy AA6111 at lower temperature in the roll gap. the Al-5Ti-0.2B master alloy, the standard grain refiner
The cast strips showed some typical microstructural fea- in aluminum industry for the manufacturing of alumi-
tures of semisolid casting. They could cast with much num foil products. The addition of grain refiner to the
higher casting speed upto 60 m=min. However, the strips melt was carried out under control conditions before
were strongly influenced by two different roll diameters feeding to TRC processing. The microstructural
across the thickness. Microstructure had distinct fea- comparison was carried out by Alcoa Cold Finger test
tures on the strip surface close to smaller roll. One of samples.
the main ideas to shift the macrosegregation away from One of the major differences in TRC and normal
centerline of strip did work but could not be eliminated. casting processes is the difference in cooling rates and
In a similar TRC setup later, Haga et al. [15] cast hyper- undercooling generated due to the variation in cooling
eutectic Al-Si alloys containing 16–20% Si. The size of rates. Thus, more constitutionally undercooling related
the primary Si was smaller than 0.1 mm. The strip growth is seen in normal casting and it is thermal in
showed good ductility to be amenable to cold rolling. TRC. Thus is the above study with Al-5Ti-0.2B grain
However, the microstructure was not uniform across refiner was more effective than Al-3Ti-0.15C grain
the thickness, and most of the primary Si existed in refiner. The latter is more suitable for normal castings.
center area of thickness. These grain refiners require large undercooling to
In the MDTRC process, Haga et al. [19–22] tried to become effective. However higher addition of Al-3Ti-
cast the alloy such that semisolid layer of the alloy is 0.15C master alloy would be required which is not desir-
dragged into the roll gap. The rolls were specially able. Out of the reasons for TiC being less effective with
designed to have low separating force. Since the rolls time is its higher sinking rate due to density difference
were made of copper, it was claimed that high cooling with aluminum and consequent settling process.
rates were maintained. The microstructures presented The effect of the grain refiners of the Al–Ti–B and Al–
for AA5182 alloy strips at different roll speeds show Ti–C on as-cast structure has also been investigated in
progressively finer structure with increasing the roll conjunction with variation in strip thickness and roll
speed [19]. The window of variables to achieve sound casting speed by Yun et al. [57]. They found that Al–
strips has been provided but rationalization is mainly Ti–C have slightly better grain refining ability for the
on experience basis. No comments on macrosegregation roll cast AA8111 alloy at higher casting speeds and
have been given. As presented earlier in processing sec- thinner cast gauges.
tion, Haga et al. [17] in the modified process where a
cooling slope was used to lower superheat with the Melt Shearing
intention of casting the strip in the semisolid state, some In industrial practice, Al–Ti–B master alloys are com-
remarkable processing advantages were shown. The monly used as grain refiners for wrought Al alloys. How-
alloys cast were AA1050, AA3004, AA5083, AA5182, ever, only less than 1% of the added TiB2 particles are
AA6063, and A356. Special advantages were claimed active for nucleation of a-Al grains [58]. TiB2 particles
658 N. S. BAREKAR AND B. K. DHINDAW
FIGURE 7.—As cast microstructures of Al-5Mg alloy through thickness (5 mm) from longitudinal cross sections of strips produced by (a) TRC and
(b) MC–TRC process. MC–TRC strip shows fine grain structure and minimal centerline segregation [68].
Downloaded by [New York University] at 16:01 06 January 2015
have a tendency to agglomerate and are detrimental in segregation is visible in the melt-conditioned TRC sheets
the final microstructure [59], particularly for products as compared to the TRC sheet cast under identical
intended for rolling. Therefore, more recently, it was thermal conditions. Another issue reported was uniform
found that by shearing the liquid metal before casting distribution of intermetallics in MC–TRC sheets as com-
using the melt conditioning through intensive melt pared to TRC sheets. Theoretically, if the solute content
shearing, it is possible to achieve a fine and uniform is more than 3%, it is difficult to cast a strip without any
microstructure and chemical composition [60]. It is well segregation. Practically, the solute content limit is 2.5%
known that oxides bifilms [61] are present in Al alloy [6]. It has been demonstrated that the melt conditioning
melt. It has been suggested [62] that these oxides can increases the range of alloys (solute content up to 7%)
act as potent nuclei if present in sufficient numbers, that can be twin-roll cast without severe segregation
favorable particle size, and size distribution. Also the [67]. Figure 7 shows the microstructure from the longi-
lattice structure of some of these oxides seems to favor tudinal cross section of as-cast Al-5Mg binary alloys
epitaxial growth of a-Al grains [63]. Thus, if these parti- strips produced by TRC and MC–TRC process.
cles are in Al alloy melt, they can act as potent nuclei, and Enhanced nucleation by melt conditioning favors the
no external addition of nucleating agent is required. advance of an equiaxed solidification front resulting in
At Brunel Centre for Advanced Solidification finer microstructure and minimum centerline segre-
Technology, melt conditioning techniques have been gation. The measured hardness values of the conven-
developed where oxides present in aluminum alloys are tional TRC and MC–TRC of Al–Mg alloys strips are
dispersed inside the melt by high shear treatment of presented in Table 1. The refinement in the grain struc-
the melt. Two types of setup are used for intensive shear- ture accounts for an increase in the hardness values of
ing of the melt. In one case the melt is fed into twin MC–TRC samples.
screws [64], and in another it is sheared by a rotor stator
device [65]. The melt is intensively sheared whereby the
oxides present are finely dispersed in the melt and are SUMMARY OF PROCESS–STRUCTURE RELATIONSHIP
expected to be transformed into potent nuclei capable The outcomes from a range of TRC experiments and
of refining the microstructure. In an earlier attempt, twin modeling have been compiled in this section. Table 2
screw-sheared melt of alloy AA5754 was subjected to provides a summary of the major reported findings of
TRC [66]. The strips were cast using water-cooled steel the effect of process variables on sump depth and center-
rolls. Significant differences in microstructure and cen- line segregation.
terline segregation were reported. Very little centerline There was no significant advantage of using unequal
diameter rolls and higher roll diameter in terms of mech-
TABLE 1.—Vickers hardness (HV5) of TRC and MC–TRC Al–Mg binary
anical performance of TRC sheets and amount of segre-
alloys strips [68]. gation in the strip. It is difficult to conclude the distinct
advantages of semisolid TRC in eliminating the center-
Alloy TRC MC–TRC line segregation (particularly, deformation segregates
and banded structure) even though the globular struc-
Al-3Mg 53.73 2.8 63.5 3.5
ture is achieved. Hence, more experimental work is
Al-5Mg 67.17 2.3 72.38 3.5
Al-7Mg 83.06 3.9 84.28 3.5
required to understand the segregation behavior as
affected by process parameters. TRC of large freezing
TWIN-ROLL CASTING OF ALUMINUM ALLOYS 659
TABLE 2.—Influence of casting parameters and thermophysical properties and roll pressure are the critical parameters in deciding
of melt on sump depth and centerline segregation during TRC. the quality of the strip. Much work has been focused
Effect on sump depth (D), centerline
on understanding the defect formations in the strips dur-
Process variables segregation (Sc) ing TRC, however not many attempts have been made
to eliminate the internal defects, in particular, centerline
Casting velocity (Vs) Vs " D " S c " segregation. Further work is necessary in this direction
Melt superheat (4Ts) 4Ts " D " Sc "
to extend the range of alloys that can be twin-roll cast
Setback (d) d " D " Sc "
Freezing range (4T) 4T " D " Sc "
with tolerable amount of segregation for a particular
Roll-strip interface heat transfer q # D " Sc " end application.
twin roll caster. Journal of Materials Processing Technology 34. Miyazawa, K.; Szekely, J. A mathematical model of the splat
2007, 191, 238–241. cooling process using the twin-roll technique. Metallurgical
16. Haga, T.; Tkahashi, K.; Ikawa, M.; Watari, H. Twin roll cast- and Materials Transaction A 1981, 12 (6), 1047–1057.
ing of aluminum alloy strips. Journal of Materials Processing 35. Saxena, A.; Sahai, Y. Modelling of fluid flow and heat
Technology 2004, 153–154, 42–47. transfer in twin roll casting of aluminum alloys. Materials
17. Haga, T.; Tkahashi, K.; Ikawa, M.; Watari, H. A vertical- Transaction 2002, 43 (2), 206–213.
type twin roll caster for aluminum alloy strips. Journal of 36. Gupta, M.; Cook, D.P.; Sahai, J. Strip casting of aluminum
Materials Processing Technology 2003, 140, 610–615. using twin roll casters. In Light Metals 1999, TMS; Eckert,
18. Haga, T.; Suzuki, S. Study on high-speed twin-roll caster for C.E., Ed., 1999; 925–930.
aluminum alloys. Journal of Materials Processing Technology 37. Bradbury, P. A mathematical model for the twin roll casting
2003, 143–144, 895–900. process, D. Phil. Thesis, University of Oxford, Department of
19. Haga, T.; Suzuki, S. Roll casting of aluminum alloy strip by Materials, December 1994.
melt drag twin roll caster. Journal of Materials Processing 38. Bradbury, P.J.; Hunt, J.D. A coupled fluid flow, deformation
Technology 2001, 118, 165–168. and heat transfer model for a twin roll caster. In Proceedings
20. Haga, T. Semi-solid roll casting of aluminum alloy strip by of the 7th Conference on Modelling of Casting, Welding and
melt drag twin roll caster. Journal of Materials Processing Advanced Solidification Processes, TMS 1995; Cross, M.,
Technology 2001, 111, 64–68. Campbell, J., Eds., London, UK, 1995; 739–746.
21. Haga, T. Semisolid strip casting using a twin roll caster 39. Satioglu, K.; Thevoz, P. Simulation of solidification process
equipped with a cooling slope. Journal of Materials Processing used in twin roll casting. In Light Metals 2003; Crepeau,
P.N., Ed., 2003; 747–751.
Downloaded by [New York University] at 16:01 06 January 2015
Al–Mg–Si alloy produced by twin roll casting process. AZ91D alloy sheets produced by melt conditioned twin
Materials and Design 2011, 32, 4603–4607. roll casting. Materials Science and Technology 2009, 25 (5),
51. Birol, Y. Analysis of macro segregation in twin-roll cast 599–606.
aluminum strips via solidification curves. Journal of Alloys 61. Campbell, J. An overview of the effects of bifilms on the
and Compound 2009, 486, 168–172. structure and properties of cast alloys. Metallurgical and
52. Birol, Y. Centerline segregation in a twin-roll cast AA 8011 Materials Transaction B 2006, 37 (6), 857–863.
alloy. Aluminum 1998, 74, 318–321. 62. Wang, Y.; Li, H.-T.; Fan, Z. Oxidation of aluminum alloy
53. Berg, B.S.; Hansen, V.; Zagierski, P.T.; Nedreberg, M.L.; melt and inoculation by oxide particles. Transactions of Indian
Olsen, A.; Gjcnnes, J. Gauge reduction in twin-roll casting Institute of Metals 2012, 65 (6), 653–661.
of an AA5052 aluminum alloy: the effects on microstructure. 63. Fan, Z. An epitaxial model for heterogeneous nucleation on
Journal of Materials Processing Technology 1995, 53, potent substrates. Metallurgical and Materials Transaction A
65–74. 2013, 44, 1409–1418.
54. Okayasu, M.; Sato R.; Takasu, S.; Niikura, A.; Shiraishi, T. 64. Barekar, N.S.; Hari Babu, N.; Dhindaw, B.K.; Fan, Z. Effect
Mechanical properties of Al Si Cu alloys produced by the of intensive shearing on morphology of primary silicon and
twin rolled continuous casting process. Material Science and properties of hypereutectic Al–Si alloy. Materials Science
Engineering A 2012, 534, 614–623. and Technology 2010, 26 (8), 975–980.
55. Slamova, M.; Karlık, M.; Robaut, F.; Slama, P.; Veron, M. 65. Patel, J.; Zuo, Y.; Fan, Z. Liquid metal engineering by
Differences in microstructure and texture of Al–Mg sheets application of intensive melt shearing. In Proceedings of the
produced by twin-roll continuous casting and by direct-chill 2013 International Symposium on Liquid Metal Processing
Downloaded by [New York University] at 16:01 06 January 2015
casting. Materials Characterisation 2003, 49, 231–240. and Casting, TMS 2013; Krane, M.J.M., Jardy, A.,
56. Birol, Y. The performance of Al–Ti–C grain refiners in Williamson, R.L., Beaman, J.J., Eds., 2013; pp 291–299.
twin-roll casting of aluminum foilstock. Journal of Alloys 66. Kumar, S.; Hari Babu, N.; Scamans, G.M.; Fan, Z. Micro-
and Compound 2007, 430, 179–187. structural evaluation of a melt conditioned twin roll cast
57. Yun, M.; Lockyer, S.A.; Hunt, J.D.; Cook, R.; Bristow, D.J. Al-Mg alloy. Materials Science and Technology 2011, 27
A study of TiCAl 315TM grain refinement in roll cast AA8111 (12), 1833–1839.
aluminum alloy. Light Metals 2000, 857–864. 67. Barekar, N.S.; Das, S.; Fan, Z. Microstructural evaluation
58. Greer, A.L. Control of grain size in solidification. In Solidifi- during melt conditioned twin roll casting (MC-TRC) of al-mg
cation and Casting; O’Reilly, K., Cantor, B., Eds.; Institute of binary alloys. In Proceedings of the 6th International Confer-
Physics: London, UK, 2003; 199–214. ence on Solidification and Gravity, Hungary, Sept 2–5, 2013,
59. Murty, B.S.; Kori, S.A.; Chakraborty, M. Grain refinement Accepted.
of aluminum and its alloys by heterogeneous nucleation and 68. Barekar, N.S.; Das, S.; Fan, Z. Melt conditioned twin roll
alloying. International Materials Review 2002, 47 (1), 3–29. casting (MC-TRC) for aluminium alloys. In Proceedings of
60. Bian, Z.; Bayandorian, I.; Zhang, H.; Scamans, G.M.; Fan, Z. 14th International Conference on Aluminium Alloys, Norway,
Extremely fine and uniform microstructure of magnesium Jun 15–19, 2014, Accepted.