Journal of Materials Processing Tech.

294 (2021) 117146

Contents lists available at ScienceDirect

Journal of Materials Processing Tech.

journal homepage: www.elsevier.com/locate/jmatprotec

Improved degassing efficiency and mechanical properties of A356

aluminium alloy castings by high shear melt conditioning
(HSMC) technology
Jaime Lazaro-Nebreda *, Jayesh B. Patel , Zhongyun Fan
Brunel Centre for Advanced Solidification Technology, Brunel University London, Uxbridge, UB8 3PH, United Kingdom

A R T I C L E I N F O A B S T R A C T

Associate Editor: A Luo The degassing of aluminium alloy melts is a crucial step in the production of high-quality casting products, as the
presence of dissolved hydrogen and oxide bi-films is detrimental to the mechanical properties. Current rotary
Keywords: degassing techniques are effective, but they lack efficiency because of the high gas flow and long processing
Aluminium alloy times required. This study aims to solve this problem by presenting an innovative rotor-stator degassing tech­
High shear melt conditioning
nology, that combines controlled inert gas injection with intensive melt shearing. It has been applied to the liquid
Degassing
metal treatment of an aluminium cast alloy to evaluate the effect on melt cleanliness, casting integrity and
Oxide bi-film
Mechanical properties mechanical properties. The optimum conditions for an efficient bubble dispersion have been obtained by water
Casting defects modelling. The melt quality during and after degassing has been assessed by in-situ measurement of hydrogen
concentration and by reduced pressure test sampling for oxide bi-films and porosity content evaluation. This new
technology is faster, requires less gas flow consumption and produces higher melt quality than the existing
degassing techniques, due to a characteristic combination of distributive and dispersive mixing flow. In addition,
re-gassing is minimised, maintaining a high melt quality for longer time after processing. This results in castings
with less defects and better mechanical properties. The improved degassing efficiency of this technology makes it
an excellent alternative in industry to increase melt quality and casting productivity.

1. Introduction moisture, as described in Eq. 1, producing atomic hydrogen, that quickly

dissolves in the melt, and alumina that is deposited in the surface as a
Aluminium alloys are widely used in various industrial sectors for the thin layer. When the melt is disturbed during melt handling, the layer
purpose of weight saving and because of the excellent combination of breaks and folds over, forming a so called bi-film oxide that becomes
properties, as reported by Polmear (1995) However, the presence of entrained inside the melt.
porosity in the final casting components is quite common and has a
detrimental effect on the mechanical properties, especially the elonga­ 2Al + 3H2O → Al2O3 + 6H (1)
tion and tensile strength, as reported by Cáceres and Selling (1996) and Hydrogen has been normally considered the origin for gas porosity in
the fatigue behaviour, as reported by Ammar et al. (2008). Porosity aluminium, because of the much lower solubility in solid than in liquid.
originates during the solidification stage and once formed it is difficult During solidification, the excess of hydrogen is rejected from solution,
to remove either by heat treatments or deformation processing, as re­ and recombines as molecular gas in the interdendritic regions. Campbell
ported by Chaijaruwanich et al. (2006). It is therefore necessary to (2003) suggested, however, that the oxide bi-films play a more impor­
ensure a good melt quality before casting by eliminating, or at least tant role in the formation of gas porosity. Each oxide bi-film present two
reducing, the causes of porosity. sides, the outer is the wet side in contact with the liquid and the inner is
Campbell (2003) mentioned two major factors that determine the the dry side in contact with entrapped air. Dispinar and Campbell (2004)
quality of an aluminium melt. One is the presence of oxide films with described that during the solidification, the bi-film unfolds and any
poor wettability and the other is the dissolved hydrogen. Both are hydrogen in excess in the melt automatically diffuses into the gap within
consequence of reaction of the liquid aluminium with the ambient water the bi-film sides. Tiryakioglu (2020) has recently shown that hydrogen

* Corresponding author.
E-mail address: jaime.lazaronebreda@brunel.ac.uk (J. Lazaro-Nebreda).

https://doi.org/10.1016/j.jmatprotec.2021.117146
Received 16 December 2020; Received in revised form 9 March 2021; Accepted 12 March 2021
Available online 15 March 2021
0924-0136/© 2021 The Authors. Published by Elsevier B.V. This is an open access article under the CC BY license (http://creativecommons.org/licenses/by/4.0/).
J. Lazaro-Nebreda et al. Journal of Materials Processing Tech. 294 (2021) 117146

gas porosity can not nucleate in liquid aluminium and therefore the previous results have been included in this paper (section 3.1) with a
initiator of porosity is the bi-film, with the hydrogen just acting as an more detailed discussion. However, no mechanical properties were re­
agent that makes defects more visible by contributing to the pore ported in any of those previous studies. Recently, Zhang et al. (2018)
expansion. Therefore, not only the hydrogen but, more importantly, the found that the use of high shear melt conditioning after degassing re­
bi-films should be removed during the degassing process before casting duces the variation of the mechanical properties of HPDC components
to reduce its potency to form porosity during solidification. based on the A380 alloy, via oxide dispersion and enhanced grain
Zhang et al. (2011) reviewed current methods for degassing molten nucleation during solidification, and Lordan et al. (2019) found that it
aluminium that are in industrial use or under research and development, also helps reducing the size and amount of porosity and defects in the
which include the injection of inert or reactive gas with lances or porous HPDC castings. However, no significant increment on the mechanical
plugs, flux addition, vacuum degassing, ultrasound degassing, spray properties was observed in these studies because of the inevitable high
degassing and rotary degassing. Among them, the most popular method level of defects originated during this casting process, as recently re­
used in industry is rotary degassing, because of its simple implementa­ ported by Zhang et al. (2020).
tion and good performance in comparison with the other techniques. In Despite the observed potential of the high shear melt conditioning
rotary degassing, the smaller and the better distributed the gas bubbles technology, the available information about the most efficient degassing
are, the higher the degassing efficiency. Warke et al. (2005a) and Warke conditions in terms of gas flow and rotor speed, is limited. Moreover, the
et al. (2005b) found that increasing rotor speed reduces the size of the full details of the effect of the melt quality after high shear degassing on
bubbles and accelerates hydrogen removal, but also causes surface tur­ the mechanical properties has not yet been evaluated. Therefore, the
bulence and a vortex, which accelerates the reabsorption of hydrogen aim of the present work is to fill this gap of knowledge and compare
(re-gassing) and the entrapment of new oxide films in the melt. There­ process efficiency with that obtained with conventional degassing
fore, the rotor speed cannot be too high in conventional rotary degass­ methods, such as tablet fluxing and rotary degassing. This includes
ing. Baffles are normally used in industry to prevent vortex appearance physical modelling in water to assess the effect of gas flow and rotation
but are only effective to certain extent and for small melts. Camacho-­ speed on bubble dispersion, the in-situ measurement of melt quality
Martínez et al. (2010) suggested that increasing gas flow rate could during and after degassing, and the evaluation of mechanical properties
decrease the vortex size by affecting the fluid flow around the impeller and integrity of components produced by gravity die casting at different
and Hernández-Hernández et al. (2015) found that increasing gas flow times after the melt treatment.
rate can also speed up the degassing process. However, Zuo et al.
(2016a,b) reported that above a certain gas flow rate the process effi­ 2. Experimental procedure
ciency decreases as bubbles escape too fast to the surface. This reduces
the collection of hydrogen and can also cause severe surface turbulences. 2.1. Material melting and casting process
Therefore, the gas flow rate is also limited in conventional rotary
degassing in terms of process efficiency. The secondary aluminium cast alloy A356, with composition Al-
Current research on rotary degassing is focused on computer 6.87Si-0.36Mg-0.12Fe (wt.%), was used in this study. The alloy was
modelling to assist process condition optimization and on the re-design melted in charges of 10 kg ingots in three clay graphite crucibles (sal­
of the rotary impeller. In that sense, Hernández-Hernández et al. (2016) amander A25; top diameter of 210 mm, bottom diameter of 155 mm and
proposed a modified impeller design that can help in better distribution height of 280 mm) in an electrical resistance furnace at a temperature of
of bubbles, as well as reducing processing time and, with it, gas con­ 700 ± 5 ◦ C. One hour after homogenisation each batch of melt was
sumption between 14 % and 34 % compared with commercial designs, degassed using a different method (section 2.2). After degassing, the
while Wan et al. (2018) recommended the variation of immersion depth melts were skimmed and held isothermally in the furnace for up to 60
of the impeller for the same purpose. Despite these efforts, rotary min. No granular flux was used to cover and protect the surface of the
degassing efficiency is still not high enough mainly due to the need for melt to allow natural re-gassing to occur during the holding period. The
relatively high gas flow rates and the long processing times of 20− 30 melt quality was assessed at different times, both during degassing and
mins for lowering the hydrogen content to industrially accepted levels. holding stages, by standard techniques (section 2.3). Additionally, ten­
Moreover, Mostafaei et al. (2016) reported that the operating window in sile bars were cast at certain times during the holding stage to evaluate
which rotary degassing produces good results is small and that a wrong the effect of each melt treatment on the mechanical properties (section
combination of process conditions may result in no improvement or 2.4) and on the as-cast microstructure (section 2.5).
even degradation in quality of castings. More recent study by Gyarmati
et al. (2020) reported that without flux addition, rotary degassing is only 2.2. Degassing methodology
effective at removing hydrogen, but not the oxide bi-films in the melt.
Regarding this issue, Fan et al. (2009) demonstrated that melt con­ The objective of the study was to evaluate the effect of the new high
ditioning by an advanced shear technology can effectively disperse shear degassing technology, but to avoid bias and for comparison pur­
oxide films into fine individual particles, and Zuo et al. (2011) reported poses, three degassing methods were used during this investigation, one
that when used in combination with Ar gas, intensive melt shearing has on each crucible. In the first crucible, degassing was done by adding 0.2
a significant degassing effect on aluminium alloys. Based on this prin­ wt.% of degassing tablet (FOSECO Nitral C19) into the melt, using a bell
ciple, Fan et al. (2011) developed a novel rotor-stator high shear degassing plunger to push the tablet to the bottom of the crucible. The
degassing technology, which allows the rotation speed to be increased plunger was withdrawn when the bubbling at the surface stopped
without vortex generation. This new technology was initially tested by completely (~5 min). The second crucible was degassed using a com­
Zuo et al. (2013) in molten A356 alloy and later by Zuo et al. (2015) and mercial rotary degasser (FOSECO, φ90 mm) for 10 min. The impeller
Kang et al. (2015) in molten A7032 alloy, finding in all cases a signifi­ was immersed near the bottom of the crucible (~50 mm) and operated
cant reduction in hydrogen and in oxide bi-film content when compared at a stirring speed of 350 rpm and an Ar flow rate of 5 L/min. The third
to conventional rotary degassing. Furthermore, they analysed the pos­ melt was processed using the innovative high shear melt conditioning
sibilities of this novel technology and discussed the potential economic (HSMC) degassing technology developed at BCAST (section 2.2.1) dur­
benefits when implemented in industry for improving the integrity of ing 10 min. The unit was immersed halfway from melt surface to the
the castings. Lazaro-Nebreda et al. (2018) continued the investigations bottom of the crucible (~125 mm) and operated at a rotor speed of 3000
on the HSMC technology for melt purification purposes and reported on rpm and Ar flow rate of 0.1 L/min. These processing parameters were
the importance of reducing the inlet gas flow in the degassing process to selected after physical modelling in water for obtaining efficient con­
achieve a better bubble dispersion and degassing efficiency. These ditions for bubble dispersion (section 2.2.2).

2
J. Lazaro-Nebreda et al. Journal of Materials Processing Tech. 294 (2021) 117146

Fig. 1. Schematic of the rotor-stator HSMC unit. (a) lateral and bottom view, (b) inside a melt illustrating the degassing procedure.

2.2.1. High shear melt conditioning (HSMC) degassing technology by an external pipe (4 mm inner diameter) placed 20–30 mm under­
The high shear melt conditioning (HSMC) technology developed at neath the unit (Fig. 1b). Camacho-Martínez et al. (2012) reported that
BCAST by Fan et.al. (2011) comprises of a rotor-stator arrangement the use of gas injection underneath the impeller with an external pipe,
made from an inert ceramic material. The HSMC model used in this instead of through the shaft, improves the bubble distribution in rotary
study, HSM-40-AL (Fig. 1a), is the same that Zuo et al. (2013) used in the degassing and allows the gas flow rate to be reduced, as required. This
initial development of the high shear degassing technology and further way, the incoming bubbles are captured by the rotor spinning, which
studies. The rotor has four blades forming a cross, with a diameter of creates a pumping action, and are then subjected to intensive shearing
29.5 mm. The stator outer and inner diameters are 42 mm and 30 mm, and dispersed into many tiny bubbles when they pass through the stator
respectively. This gives a small gap of 0.25 mm between the rotor blades openings and are distributed through the whole melt following the
and the inner wall of the stator. The stator openings are disposed as 4 macroscopic flow pattern provided by the high velocity jets.
rows of 16 circular holes of 2.5 mm diameter uniformly distributed
around the stator perimeter. 2.2.2. Physical modelling of the degassing process in water
During its operation in molten metal (Fig. 1b) the rotation speed can To visualize how the bubble dispersion occurs with the use of the
vary in the range of 1000− 10000 rpm, providing an extremely high HSMC degassing technology, controlled tests were done prior to melt
shear rate (up to 104-105 s− 1). This high shear rate is the result of the treatment using a transparent glass tank (base 210 mm) and two
melt being pumped upwards from the bottom and then squeezed in modelling HSM-40-AL units made of plastic. The tank was filled with
between the small rotor-stator gap and through the openings of the water (~250 mm height) and the unit was immersed halfway below the
stator. The liquid is projected radially, as high velocity jets, towards the liquid surface. The gas pipe end was placed 30 mm below the unit and
wall of the crucible where it splits into upper and lower flow patterns the tests were carried out by operating the unit at different rotor speeds
(dashed arrows in Fig. 1b), while more liquid is being pumped into the and under varied Ar gas flow rates. Two gas flow rates were studied, 1 L/
unit. Therefore, the technology provides not only macro-flow in a large min (high rate) and 0.1 L/min (low rate), and the rotor speed was
volume of melt for distributive mixing, but also micro-flow with inten­ increased from 1000 to 6000 rpm. The process was recorded using a
sive shearing effect near the tip of the device for dispersive mixing. digital camera.
In the HSMC degassing technology, the gas is injected into the melt Regarding the applicability of the results found in this section for the
degassing of aluminium melts, it should be noted that recent studies by
Tzanakis et al. (2017) and Yamamoto et al. (2018) have reported that
water shares similar fluid and dynamic behaviour (viscosity to density
ratio) with molten aluminium and is the most appropriate liquid to
replicate bubble dispersion and degassing performance in molten
aluminium. In the working range, above 1000 rpm, the HSMC mixer
flow is always in the turbulent regime both in water or aluminium, i.e.,
the Reynolds number is Re > 104, with it defined by Re = ρND2/μ as in
Hakansson (2018), where ρ is the fluid density and, μ is the fluid vis­
cosity, N is the rotor speed and D is the characteristics length, here the
rotor diameter.
Furthermore, the crucibles used for the aluminium melt treatment
were chosen to be of similar dimensions to the glass tank employed
during the bubble dispersion observations in water. This way the bubble
dispersion behaviour in liquid aluminium is expected to be similar to
what is observed in water and the results obtained in section 3.1 with
water modelling (rotor speed and gas flow) could be directly imple­
mented in the aluminium melt processing as shown from section 3.2.

2.3. Melt quality evaluation

Fig. 2. (a) Geometry of the standard ‘Stahl’ mould used to produce tensile
specimens according to ASTM B108 (arrows indicate the nominal flow of the The hydrogen concentration in the melts, before, during and after the
liquid metal); (b) tensile specimen geometry and dimensions (dotted line degassing melt treatments, was directly measured by using a FOSECO
indicate region for pre-test microstructure evaluation). Dimensions in mm. ALSPEK-H probe. The probe was immersed at a depth of between 150

3
J. Lazaro-Nebreda et al. Journal of Materials Processing Tech. 294 (2021) 117146

Fig. 3. Physical modelling of the HSMC degassing process in water, with variation of rotor speed and Ar gas flow rate. (a–d) 1 L/min (e-h) 0.1 L/min. Figure adapted
from Lazaro-Nebreda et al. (2018).

and 200 mm in the melt. The readings were given in cm3/100 g with real 2003). The yield strength at 0.2 % offset (Ys), the ultimate tensile
time data. strength (UTS) and the elongation at break (El%) were evaluated from
The reduced pressure test (RPT) was also considered in this study. the stress-strain curves. Each data and standard deviation reported in the
The method involves solidifying the melt into two conical steel cups, one results was based on properties obtained from the 4 samples cast each
in air (atmospheric pressure) and the other under partial vacuum (80 time.
mbar or 8 kPa). The cups were always preheated above 200 ◦ C prior to
each test to prevent rapid solidification and ensure reliable results, as 2.5. Microstructure characterization
recommended by Tan et al. (2011). The melt quality was accessed by
using the density index (DI), which was calculated using Eq. 2, where Selected tensile specimens were examined for microstructure and
Dair and Dvac are the density of samples solidified in air and under porosity, both before and after tensile testing, to better understand the
vacuum, respectively. nature of the mechanical failure. For the pre-test characterization, a slice
was cut from the top part of the sample (location marked with a dotted
DI = (Dair – Dvac) / Dair (2)
line in Fig. 2b). The tested specimens were examined at the location of
After calculating the density index, selected samples were sectioned, fracture. Samples were prepared by standard metallographic tech­
ground and polished to evaluate the porosity size and distribution by niques. Microstructure images were acquired using a Zeiss optical mi­
using the image analysis software ImageJ. croscope and analysed using ImageJ software.

3. Results and discussion

2.4. Tensile property evaluation
3.1. Effect of HSMC on bubble dispersion by water modelling
Following each degassing treatment, sets of tensile bars were cast at
certain times (10 min, 30 min and 50 min) during the isothermal holding Fig. 3 shows a selection of the most representative snap-shots from
stage by manually pouring with a ladle into an ASTM B108 standard the videos recorded during the water experiments under different con­
permanent steel mould (ASTM, 2019). Fig. 2a shows the geometry of the ditions. Dashed arrows have been included in the images to indicate the
mould used, with arrows indicating the nominal flow of liquid metal. In bubble flow path and allow for a better observation of the region
total, 12 tensile specimens were produced from each melt (4 at each affected by the dispersed bubbles. For the higher gas flow rate (1 L/min)
holding time), with the key dimensions shown in Fig. 2b. Care was taken and 2000 rpm, it can be observed in Fig. 3a how the large bubbles
to pour the material under the same conditions each time, by keeping (10− 20 mm) leaving the incoming pipe were captured by the rotor-
the mould preheated inside a furnace at 460 ◦ C before each casting, as stator unit and came out of the stator holes as high velocity air-water
reported by Dong et al. (2018). jets containing smaller size bubbles (< 1 mm). These bubbles then fol­
After being kept at ambient temperature for 24 h, the tensile speci­ lowed a parabolic trajectory and travelled fast upwards and disturbed
mens were tested in the as-cast condition at room temperature using an the surface similarly to what happens in commercial rotary degassing.
Instron 5500 universal electromechanical testing system equipped with No vortex formation is observed, even though the rotor speed is much
a 50 kN load cell, in accordance with ASTM standard E8-03 (ASTM, higher than what is normally used with conventional rotary degassers

4
J. Lazaro-Nebreda et al. Journal of Materials Processing Tech. 294 (2021) 117146

rate, G, through the expression DαG1/3 and only weakly affected by the
nozzle diameter. Therefore, the lower the incoming flow rate, the
smaller the initial bubbles to be dispersed. In the present study the gas
flow of 1 L/min injected bubbles of about 15 mm (Fig. 3a), similarly
reported by Zuo et al. (2016a,b), and the gas flow of 0.1 L/min injected
bubbles of approximately 6 mm diameter, as shown in Fig. 3e-h. In the
same way, Kang et al. (2015) reported injected bubbles of 10 mm
diameter when using an inlet gas flow of 0.3 L/min. These previous
studies on the HSMC unit also reported that the sharp decrease of the
bubble size occurs more strongly for rotor speeds up to 3000 rpm, with
almost no change for higher rotor speeds. Therefore, it was expected that
a lower incoming gas flow rate would help to produce much smaller
dispersed bubbles for the same rotor speed compared with a higher flow
rate.
This can be appreciated in Fig. 3e-h, for a flow rate of 0.1 L/min,
where the bubbles coming out of the tube and out of the stator holes
were much smaller compared to the case of 1 L/min (Fig. 3a-d). The first
Fig. 4. Volume fraction of water affected by bubbles as a function of rotor
consequence of this is that bubble buoyancy was much slower and the
speed and Ar flow rate. The 0.3 L/min line is reproduced from data in Kang
et al. (2015).
instabilities due to bubble rupture at surface were minimized at all rotor
speeds. This is more clearly observed in the video recordings. Secondly,
at same rotor speed of 2000 rpm (Fig. 3f) the affected region by
(~500 rpm). With an increase of the rotor speed to 3000 rpm the bubble
dispersed bubbles was larger in comparison with the one observed for
size decreased, and the affected region became larger (Fig. 3b). Bubble
the higher flow rate (Fig. 3a). This is because the water jets can carry the
buoyancy also decreased and with it the effect on surface instability.
smaller bubbles further. Thirdly, the fine bubble jets reached the wall of
However, no bubbles were observed below the unit. As the rotor speed
the tank when the rotor speed was at around 2500 rpm (Fig. 3g), with
got closer to 4000 rpm (Fig. 3c) the bubble jet flow hit the wall of the
consequently earlier presence of dispersed bubbles in most of the vol­
tank and divided by following both upward and downward paths, the
ume below the unit, and a clear full bubble recirculation could be
latter carrying the bubbles with it towards the region below the unit.
observed when the rotor speed was 3000 rpm (Fig. 3h) and above.
Increasing the rotor speed to 5000 rpm enhanced the presence of the
Fig. 4 gives the quantitative evaluation of the volume fraction of
bottom dispersed bubbles towards the base of the water tank (Fig. 3d),
water in the tank showing presence of dispersed bubbles as a function of
thus promoting their full recirculation when they are captured again and
rotor speed for the two gas flow rates studied. Data from 0.3 L/min flow
re-dispersed into the high shear flow. Zuo et al. (2016a,b) had reported
dispersion reported by Kang et al. (2015) has been included for a better
in a previous analysis of the HSMC unit, in a similar water modelling
comparison of the results. A dashed line at 50 % has been drawn to
setup, that a rotor speed of 6000 rpm is recommended for effectively
highlight the moment the bubble jets hit the wall. As the unit was
dispersing the bubbles in full recirculation when the incoming Ar gas
immersed halfway inside the water tank, that point marks the transition
flow is set to 1 L/min− 1. Similarly, Kang et al. (2015) had reported on
from observing bubbles dispersed just above the stator holes and the
bubble dispersion in water by HSMC but using a lower gas flow rate of
beginning of the regime with dispersed jets both above and below the
0.3 L/min and found that this minimum rotor speed for reaching full
mixer enhancing the bubble recirculation flow. This line crosses with the
bubble recirculating regime is in the range from 4000 to 5000 rpm, and
0.1 L/min curve at 2250 rpm, with the 0.3 L/min curve at 2750 rpm and
that above 5000 rpm bubble size hardly decreases. It should be noted
it crosses at 3500 rpm for the case of the 1.0 L/min curve, thus high­
that bubbles can only survive in this recirculating motion if they are
lighting the need of higher rotor speed to achieve the same bubble
sufficiently small to stay suspended and the jet velocity is high enough
dispersion degree when increasing the gas flow rate.
for the dragging force of the downwards flow to overcome the natural
A second dashed line has been drawn, at around 95 %, to denote the
bubble buoyancy towards the surface.
moment that bubble recirculating flow reached the bottom of the tank
Scargiali et al. (2014) reported that the diameter, D, of each Ar
below the HSMC unit. Further increase of the rotor speed caused no
bubble that is released from a tube is mainly determined by the gas flow
significant change on the volume affected by the dispersed bubbles, but
rather finishing filling the corners of the tank. It can be appreciated how
a minimum rotor speed of 6000 rpm is required to reach this full bubble
recirculation regime for a gas flow of 1 L/min, as also reported by Zuo
et al. (2016a,b). This minimum rotor speed decreases to 5000 rpm for a
gas flow of 0.3 L/min, as reported by Kang et al. (2015) and is only 3500
rpm in this study for the case of using 0.1 L/min gas flow.
This is a quite significant achievement for the high shear degassing
technology as it means similar results in terms of bubble dispersion can
be obtained by reducing both injected gas flow and rotor speed. The
reduction of the inlet flow rate not only reduces process costs but also
efficiency as it uses each injected bubble to its maximum effectiveness
without losses due to bubble buoyancy. It also enhances the bubble size
reduction and the bubble recirculation flow and with it the increased
time they can remain in the fluid capturing dissolved gases before they
reach the surface.
On the other hand, the power consumption of the HSMC unit can also
be reduced. Hakansson (2018) described the power (P) transferred to
Fig. 5. Hydrogen concentration in the melt as a function of time for the three the fluid by the rotor as P = Np ρ N3 D5, where D is rotor diameter, N the
degassing methods. Degassing ends at t =10 min. Sampling of tensile bars starts rotor speed, ρ the fluid density and Np the power number. According to
at t =20 min. Hakansson (2018) the Np depends on mixer geometry but is constant

5
J. Lazaro-Nebreda et al. Journal of Materials Processing Tech. 294 (2021) 117146

with respect to rotor speed and diameter under turbulent conditions, i.e. level (0.096 cm3/100 g). With further HSMC degassing up to 10 min
Re> 104, as in the case of this study. Therefore, power consumption is the hydrogen content decreased to 0.07 cm3/100 g, which is a very low
reduced by a factor of 5 when decreasing the minimum rotor speed from level of hydrogen concentration. In comparison, the rotary degasser
6000 rpm to 3500 rpm and by a factor of 8 when using 3000 rpm instead required 10 min to reach the quasi-equilibrium level. The hydrogen
of 6000 rpm. content evolution during the degassing process, Ht, can be described by
It should be noted that these optimized parameters shown here are Eq. 3, as shown by Warke et al. (2005), where Heq denotes the equi­
particular for the water/crucible tank dimensions selected for this study. librium solubility level, H0 the initial content, Kb the mass transfer co­
However, if implemented in much larger crucibles, the earlier appear­ efficient at the melt/bubble interface, Ab the total interfacial area of the
ance of this bubble recirculation regime when using low gas flow rates, dispersed bubbles and Vm the volume of melt being processed.
would give the HSMC technology a wider range for increasing the rotor
speed without overpassing the unit working limits (<10,000 rpm), Ht = Heq + (H0 – Heq) exp(-Kb Ab t/Vm) (3)
something that would not be practical if using 1 L/min. Considering the experimental conditions, with similar H0, Heq and
Vm in the three cases, and Kb independent of the degassing conditions, as
reported by Warke et al. (2005), that leaves Ab as the key parameter to
3.2. Effect of HSMC degassing on hydrogen removal explain the differences observed in the hydrogen kinetics. Therefore, the
higher removal rate with the HSMC degassing method highlights the
Fig. 5 presents the evolution of the hydrogen concentration in the much larger interfacial area of the bubbles dispersed in comparison with
melt as a function of time for the three degassing methods used. The rotary degassing, even though the Ar gas flow used was 50 times lower.
vertical dashed line marks the start of the holding stage after degassing As shown in Fig. 3 and also reported by Kang et al. (2015) and Zuo
(t =10 min). The horizontal dotted line gives the hydrogen equilibrium et al. (2013), the bubbles dispersed by the HSMC technology can be
solubility level (0.056 cm3/100 g) in the solid alloy according to Tir­ reduced to sizes between 0.1− 0.5 mm, which is much smaller than the
yakioglu (2020). The horizontal dashed line gives the hydrogen (H) 10 mm bubble size that is usually obtained with the conventional rotary
stabilisation (quasi-equilibrium) level (0.096 cm3/100 g) due to the degassing process. For N bubbles of radius r, the total volume is V =
relative humidity (RH) in the air in the laboratory the day of the 4πr3N/3 and the total bubble surface area is A = 4πr2N. Therefore, for a
experiment, as given by the expression H = 0.1772RH + 0.0394 pro­ fixed volume of injected bubbles, the interfacial area of bubbles can be
vided by Eskin et al.(2015). expressed by A = 3 V/r, i.e. is inverse proportional to the bubble size.
All the melts started with a comparable hydrogen content (~ 0.37 ± Consequently, the interfacial area of the bubbles dispersed with the use
0.02 cm3/100 g), and the three methods all showed an initial high of the HSMC technology is much larger, at least 20–100 times, than that
hydrogen reduction rate. Despite this, the tablet method quickly reached obtained with conventional rotary degassers, explaining the faster
a steady value (~ 0.18 cm3/100 g) during the 5 min the bubbling lasted. diffusion rate of hydrogen into the bubbles during the degassing process,
It was clearly not effective at completely removing all possible hydrogen as shown in Fig. 5.
from melt as it was left far above the stabilization level. This is an ex­ In terms of the required degassing processing time with the selected
pected result since industrial practice recommends either to do tablet HSMC parameters, i.e. 0.1 L/min and 3000 rpm, the results are as ex­
degassing treatment in several steps or combined with rotary impeller to pected when compared with previous studies of the HSMC degassing
enhance the bubble distribution. When looking at the other two technology. Zuo et al. (2013) initial tests of the HSMC technology in
methods, both were effective at removing the hydrogen from the melts, A356 molten aluminium alloy recommended processing times between
reaching values below the quasi-equilibrium. Due to this they later show 1–2 min when using a rotor speed of 6000 rpm and 1 L/min gas flow rate
a slow but constant re-gassing towards the quasi-equilibrium level to fully degas a similar melt volume than the one used in this study. It is
during the holding stage. In contrast, tablet degassing did not experience known, as described by Hakansson (2018), that the net flow rate passing
re-gassing as it did not reach that low level. It should be noted that re- through the stator holes is proportional to the rotor speed. For a fixed
gassing is a slow process and in the present study the melts did not melt volume, a decrease in rotor speed from 6000 rpm to 3000 rpm
exhibit clear re-gassing right after gas injection stopped but rather when would reduce to half the flow rate and double the time required for
the surface was skimmed off and disturbed for sampling the tensile bars. processing the melt through the unit. In that way, Dybalska et al. (2017)
In terms of degassing kinetics, the HSMC degassing was faster than reported a processing time of 4 min when using a rotor speed of 3000
rotary degassing, only requiring 5 min to reach the quasi-equilibrium rpm to ensure effective particle dispersion in an alloy melt of similar
volume, which is about double the time required by Zuo et al. (2013),
and similar to the required time in the present study to lower the
hydrogen content down to the quasi-equilibrium level.
After degassing, both rotary and HSMC degassed melts exhibit a
quite stable but still decreasing hydrogen content during the initial 10
min, with final values of 0.08 cm3/100 g for rotary degassing and 0.06
mL/100 g for HSMC degassing. This further reduction is associated to
the dispersed bubbles in the melt when the rotor stops. The remaining
tiny bubbles require some time to reach the surface during which they
join up with other bubbles on the way up. As HSMC bubbles are much
smaller than rotary dispersed bubbles the resident time in the melt is
higher and with it the chances to collect and carry more hydrogen out of
the melt, explaining the final lower hydrogen content.
From the moment the melts were skimmed, and the sampling of the
tensile bars was started, the disturbance of the protective surface oxide
layer caused the hydrogen level to start increasing again (natural re-
gassing). However, for the HSMC degassed melt, the levels remained
lower, even after a holding period of 50 min. This can be also explained
Fig. 6. Density index as a function of time during degassing and isothermal by the remaining dispersed bubbles in the melt when degassing stops.
holding for the different degassing methods. Degassing ends at t =10 min. The floating velocity, Vf, towards the surface can be expressed by the
Sampling of tensile bars starts at t =20 min.

6
J. Lazaro-Nebreda et al. Journal of Materials Processing Tech. 294 (2021) 117146

index values for the HSMC melt are significantly below the ones ob­
tained for the other two degassing techniques. The tablet method always
shows density index above 6%, highlighting its ineffectiveness at
degassing. For rotary and HSMC techniques, the density index falls
below the industrially accepted 3% value after the 10 min of degassing,
although with HSMC values decreasing faster and towards lower final
values, below 1%.
The differences between the methods are also clear during the
isothermal holding stage in favour of the HSMC degassing. For the first
10 min after degassing the density index is quite constant in the three
cases. Once tensile bars sampling started at t =20 min, a significant
increase in the density index was observed for the tablet and rotary
degassed melts, while for the HSMC degassed melt the density index
remained below 1.5 % for up to 50 min after degassing and sampling.
This slow re-gassing can be as well understood in terms of the small,
dispersed bubbles left in the melt when the HSMC unit stops. These tiny
bubbles not only keep collecting hydrogen on their way up, but also act
Fig. 7. Correlation between density index (DI) and hydrogen (H) content in the as a kind of deep bed filtering system capturing the oxide particles or bi-
melt for the three degassing methods used in this study. films entrapped during sampling in the large bubble interfacial area and
migrating them to the dross layer at the surface. Zuo et al. (2013) HSMC
Stokes law as in Eq. 4, where g is the gravitational acceleration, ρAr is the degassing study using a gas flow of 1 L/min and 6000 rpm rotor speed
density of Ar, ρ melt the density of the Al melt, φ the bubble diameter and reported also reduced re-gassing during the holding stage but up to a
η the melt viscosity. maximum of 15− 20 min after degassing, after which density index
quickly started increasing way above the 3% limit. This previous result
Vf = g(ρAr - ρmelt)φ2/(18η) (4) highlights again the importance of reducing the incoming gas flow rate
to obtain even smaller and better dispersed bubbles with the HSMC
The floating speed decreases with reducing the bubble size and
degassing technology. This way they move slowly towards the surface
therefore the finer bubbles obtained by the HSMC unit and low gas flow
(Eq.4), while offering a more complicated path for the oxide particles to
rate can remain for longer in the melt. Not only they can continue
pass through, thus enhancing their attachment to the bubble surface. In
capturing hydrogen for a longer period on their way towards the melt
the present study with lower rotor speed and lower Ar gas flow, there
surface but can also hinder the re-gassing by preventing fresh hydrogen
was no re-gassing observed up to 50 min after HSMC degassing, which is
dissolving into the melt.
a significant improvement for the HSMC technology. Similar low density
index and minimal re-gassing for long holding periods when using low
3.3. Effect of HSMC degassing on Density Index reduction gas flow rate have been also reported by Lordan et al. (2019) on larger
melts, highlighting the reproducibility of the HSMC degassing results.
The results from density index evaluation during and after the That way, and provided the surface is not intensively disturbed, the
degassing treatments are presented in Fig. 6. In this case the differences processed melt by the HSMC technology can rest for longer after
between degassing methods are clearer than by just using the hydrogen degassing with the guarantee that neither hydrogen content nor density
content assessment, especially during the holding stage. The density

Fig. 8. Porosity in the RPT samples solidified under partial vacuum. (a) before degassing; and during degassing: (b) rotary, 2 min; (c) rotary, 5 min; (d) rotary, 10
min; (e) HSMC, 2 min; (f) HSMC, 5 min.

7
J. Lazaro-Nebreda et al. Journal of Materials Processing Tech. 294 (2021) 117146

Dispinar and Campbell (2004) described, the density index refers to the
potential porosity to be originated in the castings and therefore depends
not only on the hydrogen content necessary for the pore expansion but
also very sensitively the oxide bi-film content that initiate the pores. This
can be well appreciated in Fig. 7. For the tablet-degassed melt the
hydrogen content remains almost constant, but the density index ranges
between 6% and 9%. On the other hand, when the hydrogen level is
below 0.1 cm3/100 g, i.e. the density index for rotary degassed melt
varies from 2.3%–4.5%, while for the HSMC degassed melt it varies only
between 0.62 % and 1.74 %. Therefore, the rotary degassed melt density
index values are much higher and scattered than obtained after HSMC
degassing. This variability in the density index regardless of the
hydrogen was also reported by Dispinar et al. (2010) and is caused by
the dissimilar oxide bi-film content in the melts after degassing and the
way each method deals with the oxide bi-film removal. To better un­
derstand this, Fig. 8 shows the vertically sectioned RPT samples solidi­
Fig. 9. Pore size distribution of the RPT samples solidified under vacuum fied under partial vacuum at different times during degassing for the
during degassing. rotary and HSMC degassing methods, which help visualizing the dif­
ferences in the melt processing. Fig. 9 presents the pore size distributions
for the RPT samples shown in Fig. 8 and Table 1 gives the average and
Table 1
maximum pore length in each case, as well as the density of pores, which
Porosity analysis on the RPT samples solidified under vacuum during degassing.
is an indicator of the size and level of bi-films present in the melt, as
Average pore size (μm) Max. pore size (μm) pores/cm2 reported by Dispinar and Campbell (2004). In addition, Fig. 10 presents
Before degassing* 652 1200 37 the micrographs of the dross skimmed off after degassing and prior
Rotary 2 min 474 1075 35 sampling by the rotary and HSMC degassing methods, which show the
Rotary 5 min 328 825 24 type of oxides that are being removed from the melt in each case.
Rotary 10 min 339 800 22
HSMC 2 min 213 625 94
Before degassing, the RPT sample solidified under partial vacuum
HSMC 5 min 103 375 13 (Fig. 8a) exhibits the typical convex top surface caused by expansion due
HSMC 10 min 94 250 11 to porosity formation during solidification. The large pores follow a
*
Bimodal distribution. bimodal size distribution, with peaks at around 420 μm and 850 μm and
an average pore size of 652 μm. The maximum pore size is 1200 μm and
the pore density is 37 pores/cm2. With the application of Rotary
index will increase rapidly and will not require covering fluxes as it is
degassing the top surface of the RPT samples transforms from convex to
common practice after rotary degassing.
concave shape as processing time increases (Fig. 8b-d), because of the
removal of hydrogen and oxide bi-films that promote the porosity
3.4. Relationship between Density Index, hydrogen content and RPT
expansion during solidification. Pore density decreases fast towards
porosity
almost half of the initial value in 5 min and the pore size distribution
moves towards lower values, acquiring a unimodal shape. This high­
To better understand the results presented in the previous sections, a
lights that the larger bi-films are removed from the melt faster than the
correlation graph between the hydrogen measurements (Fig. 5) and the
smaller ones by the rotary degassing method, something that agrees
calculated density index (Fig. 6) is given in Fig. 7. The data exhibit a
with previous findings by Warke et al. (2005). The removal of the
good fit (R2 > 0.95) to a quadratic function (included in the graph).
smaller oxides requires of longer processing times. However, it can be
Similar good correlations were reported by Eskin et al. (2015) for the
appreciated that there is not much difference in the results from 5 min to
A356 and A380 aluminium alloys. Owing to the well-known difficulty of
10 min of processing. As it can be observed in Fig. 10a, the dross
accurately measuring the hydrogen content in aluminium alloy melts in
collected after rotary degassing consists mostly of large poorly wetted bi-
industry, this correlation reinforces the possibility of assessing the
films, those that are more easily lifted to the surface by the large bub­
hydrogen levels in the melts indirectly via the RPT sampling as it is
bles. This suggests that further rotary degassing is not fully effective at
commonly done in industry.
removing the smaller bi-films from the melt but only removing the
However, despite this good correlation it should not be forgotten that
dissolved hydrogen in the melt, as shown in Fig. 3. This causes the small
the measurement of density index considers more than just the hydrogen
natural reabsorption of hydrogen during the holding stage to produce
content and that is another reason for its extended use in industry. As

Fig. 10. Micrographs of the dross collected after melt treatment by (a) rotary and (b) HSMC degassing.

8
J. Lazaro-Nebreda et al. Journal of Materials Processing Tech. 294 (2021) 117146

Fig. 11. Representative microstructure of the tensile bars before testing for each degassing method. Samples obtained at 10 min after degassing. (a) tablet; (b) rotary
and (c) HSMC degassing.

the significant increase and variability in the DI% values, as mentioned secondary phases, which was expected as all samples were cast under
before. similar cooling conditions. The only difference between the three
On the other hand, when the HSMC degassing is applied to the melt, degassing methods was the different levels of porosity. In the sample
it can be observed how during the first 2 min of processing (Fig. 8e) the solidified after tablet degassing (Fig. 11a) pores larger than 100 μm can
number density of pores increases up to 97 cm− 2, almost 3 times the be seen, whereas for the sample obtained after rotary degassing
initial level, but the pore length decreases significantly, to almost one (Fig. 11b) the pores observed are in the range from 50 to 100 μm. On the
third of the initial size. The reason for that is that the HSMC degassing is other hand, the sample obtained after HSMC degassing (Fig. 11c)
not only dispersing the tiny bubbles in the melt but also dispersing the showed pores with size below 10 μm. With increasing holding time after
large poorly wetted oxide bi-films into smaller and better wetted bi-films degassing, no changes in microstructure were observed between the
with less potency to form large pores during solidification. With further degassing methods but the tablet and rotary degassed samples exhibited
HSMC processing up to 5 min (Fig. 8f) the small oxide bi-films are higher defect content while the HSMC tensile bars kept a low porosity
further dispersed, resulting in even smaller pores in the RPT samples. level. Similar findings were reported by Uludag et al. (2018), and are
HSMC processing up to 10 min continues dispersing the oxide bi-films, caused by the higher presence of bi-films in the melts and their expan­
resulting in reduced porosity size in RPT samples solidified under vac­ sion due to the gradual reabsorption of hydrogen during re-gassing.
uum during degassing, although the effect is less significant as the pore
size distribution or the density of pores do not vary much when
compared with HSMC for 5 min. This suggests that an increase in HSMC 3.6. Effect of HSMC degassing on tensile properties
processing above 10 min for the selected melt quantity might not
translate in further improvement of the melt quality. Fig. 10b shows The tensile properties of the tensile bars cast at different holding
how the dross collected after HSMC degassing consists of fragmented times after each degassing method are shown in Fig. 12. Each point and
and better wetted oxide films rather than the typical large bi-film layer. error bars correspond to the average and standard deviation of the re­
The poorly wetted surface inside the bi-film is therefore reduced with sults obtained from the 4 specimens cast at each condition.
the use of HSMC, and the films eventually transformed into individual The yield strength was unaffected by either the degassing method or
and well wetted particles, which are more easily captured and lifted to the holding time after degassing (Fig. 12a), with average values of 89.3
the surface by the small bubbles that the HSMC technology provides. In ± 4.1 MPa, 87.8 ± 5.1 MPa and 89.5 ± 3.9 MPa for tablet, rotary and
consequence, it is difficult for the particles to act as initiators for the HSMC degassing, respectively. Significant difference is observed, how­
porosity during solidification, regardless of the hydrogen content, ever, in the ultimate tensile strength (Fig. 12a) and in the elongation at
explaining the sharp decrease in density of pores and the reduced fracture (Fig. 12b), with HSMC degassed specimens showing the best
variability in the density index. performance with average values for UTS and elongation of 178.8 ± 6.9
MPa and 6.7 ± 0.6 %, respectively. Rotary degassed bars show average
properties of 154.1 ± 7.1 MPa and 3.3 ± 0.8 %, similar to Akhtar et al.
3.5. Effect of HSMC degassing on casting integrity before tensile testing (2009). The tablet degassed bars showed the worst results with average
values of 130.8 ± 10.3 MPa and 1.5 ± 0.4 %.
Representative microstructures of the as-cast tensile specimens, at 10 A linear decrease of UTS and elongation with holding time is
min after degassing, are given in Fig. 11 for the three degassing methods observed in all three cases (trend dashed lines and linear fitting included
considered in this study. The tensile bars exhibited similar microstruc­ in the plots for better visualization). The UTS decay is at 0.51 MPa/min
tural features in terms of dendritic arm spacing, size and distribution of for the tablet degassed bars and 0.28 MPa/min for rotary degassing. It is

Fig. 12. Tensile properties of the tensile specimens as a function of holding time after degassing. (a) yield strength (YS) and ultimate tensile strength (UTS) and (b)
elongation at break.

9
J. Lazaro-Nebreda et al. Journal of Materials Processing Tech. 294 (2021) 117146

Fig. 13. Fracture surfaces of the post-mortem tensile bars cast at 10 min after degassing. (a) tablet degassing; (b) rotary degassing; (c) HSMC degassing. (d-f) Vertical
sections of samples in (a-c).

Fig. 14. SEM micrographs showing the oxide films on the fracture surface of the tensile specimens. (a) tablet degassing; (b) rotary degassing; (c) HSMC degassing.

only 0.19 MPa/min for HSMC degassed specimens, which represents a

Table 2
reduction of 63 % compared to tablet degassing and 32 % reduction
Results of EDS analysis on the rough brighter areas of the fracture surfaces.
compared to rotary degassing. For the elongation, the decay is minimal
for tablet degassed samples, which is somehow expected considering the Element Weight % Atomic % Error %

initial low values. For rotary degassed samples the elongation decays at O 19.27 28.82 10.33
0.035 %/min, while for HSMC samples this decay is only 0.025 %/min Mg 5.21 5.13 5.29
Al 48.75 43.24 3.43
(29 % less).
Si 26.77 22.81 6.84
The tensile properties of Al-Si alloys are known to be dependent on
multiple factors such as the grain size and structure, the size and dis­
tribution of secondary phases and the size and distribution of defects. As
reported by Uludag et al. (2018) the microstructure does not vary during
the holding stage, if casting conditions are maintained, but the amount
of porosity tends to increase affecting the mechanical properties, as
shown here in Fig. 12. The reason is the remaining oxide bi-films in the
melt after degassing and their expansion during the solidification is
enhanced by the hydrogen reabsorption during holding stage. Wang
et al. (2013) reported that the employment of an in-furnace filter had a
very positive effect on reducing inclusions and oxide bi-films, therefore
decreasing casting porosity and improving the tensile properties in the
test bars. They reported values of elongation (El) from El < 4.5 % before
filtering to El > 6% after filtering. The results of the present study show
similar improvement, therefore confirming the enhanced elimination of
oxide bi-films by using the HSMC degassing technology.

3.7. Effect of HSMC degassing on fracture surface

Fig. 15. Area fraction of defects at fracture surface (f) as a function of holding
The fracture surfaces of the tested tensile bars cast at 10 min after
time after degassing.
degassing are given in Fig. 13a-c. The vertical sections of the tensile bars
at fracture locations are given in Fig. 13d-f, highlighting the differences
in defect content between the different degassing methods. The obser­ identified as oxide defects (see Table 2). Both cover most of the fracture
vation of the fracture surfaces via SEM is given in Fig. 14. Two types of surface area, which exhibits a high degree of tortuosity. For the bars
defects were clearly identified at the fractured surface of the tensile bars after rotary degassing (Figs. 13b and 14 b) the surface roughness re­
after tablet degassing (Figs. 13a and 14 a). One is the rounded voids, mains relatively high but the number of defects at the surface decreases
identified as gas porosity, and the other is the rough brighter areas, significantly, especially the gas porosity. However, the surface area still

10
J. Lazaro-Nebreda et al. Journal of Materials Processing Tech. 294 (2021) 117146

Fig. 16. Tensile properties as a function of the fraction of defects at fracture surface for (a) UTS; (b) Elongation.

exhibits a high degree of oxide defects. On the contrary, for the bars after Visualization, Supervision. Zhongyun Fan: Writing - review & editing,
HSMC degassing (Figs. 13c and 14 c) the fracture surface is smooth, with Supervision, Funding acquisition.
no visible gas porosity and smaller oxide defects.
The evolution of the area fraction of defects at the fracture surface
during the holding time for the three degassing methods is given in Declaration of Competing Interest
Fig. 15. It can be clearly appreciated how the tensile bars after HSMC
exhibit very low and constant level of defects, explaining the good The authors declare that they have no known competing financial
performance in the tensile properties. On the other hand, the tensile bars interests or personal relationships that could have appeared to influence
after tablet and rotary degassing show a constant increase of the fraction the work reported in this paper.
of defects at the fracture surface, explaining the gradual decay of the
tensile properties. Acknowledgements
A good correlation between tensile properties and the measured
fraction of porosity at the fracture surface in the tensile bars is given in The financial support from EPSRC (UK) under grant number EP/
Fig. 16. Similar findings were reported and discussed by Cáceres and N007638/1 of the Future Liquid Metal Engineering Hub is gratefully
Selling (1996) and more recently by Lordan et al. (2020), highlighting acknowledged. The high shear degassing work initiated by Dr. Yubo Zuo
the importance of good melt quality to produce castings with low is also thankfully appreciated.
amount of defects.
References
4. Conclusions
Akhtar, S., Dispinar, D., Arnberg, L., Di Sabatino, M., 2009. Effect of hydrogen content,
The innovative HSMC degassing technology developed at BCAST has melt cleanliness and solidification conditions on tensile properties of A356 alloy. Int.
J. Cast Met. Res. 22, 22–25. https://doi.org/10.1179/136404609X367245.
been investigated by water modelling and improved by determining the Ammar, H.R., Samuel, A.M., Samuel, F.H., 2008. Porosity and the fatigue behaviour of
required gas flow and rotor speed for an efficient bubble dispersion. The hypoeutectic and hypereutectic aluminum-silicon casting alloy. Int. J. Fatigue 30
key finding is that when very low gas flow is used, lower rotor speed is (6), 1024–1035. https://doi.org/10.1016/j.ijfatigue.2007.08.012.
ASTM B108 / B108M-19, 2019. Standard Specification for Aluminum-Alloy Permanent
needed to achieve an effective bubble dispersion. This has clear benefits Mold Castings. Annual Book of ASTM Standards, vol 02.02. ASTM International,
in terms of reduced processing costs and increased process efficiency. West Conshohocken, PA. http://www.astm.org/cgi-bin/resolver.cgi?B108B108M.
The improved HSMC degassing parameters have been applied to ASTM E8-03, 2003. Standard Test Methods for Tension Testing of Metallic Materials.
ASTM International, West Conshohocken, PA. http://www.astm.org/cgi-bin/resolv
aluminium melts and the results compared with conventional degassing er.cgi?E8.
methods showing better and faster removal of hydrogen and oxide bi- Cáceres, C.H., Selling, B.I., 1996. Casting defects and the tensile properties of an Al-Si-Mg
films. Rotary degassing is effective at removing hydrogen from the alloy. Mater. Sci. Eng. A 220, 109116. https://doi.org/10.1016/S0921-5093(96)
10433-0.
melt, but it is not for removing oxide bi-films. The HSMC degassing Camacho-Martínez, J.L., Ramírez-Argáez, M.A., Zenit-Camacho, R., Juárez-
technology can effectively remove both the hydrogen and the entrapped Hernández, A., Barceinas-Sánchez, J.D.O., Trápaga-Martínez, G., 2010. Physical
oxide bi-films at the same time because it can operate at higher speeds modelling of an aluminium degassing operation with rotating impellers—a
comparative hydrodynamic analysis. Mater. Manuf. Process. 25 (7), 581–591.
without causing surface turbulences. This significantly improves the
https://doi.org/10.1080/10426910903367386.
melt quality and maintains it for longer after degassing. Consequently, Camacho-Martínez, J.L., Ramírez-Argáez, M., Juárez-Hernández, A., González-Rivera, C.,
the aluminium melt processed by the HSMC degassing technology does Trápaga-Martínez, G., 2012. Novel degasification design for aluminum using an
impeller degasification water physical model. Mater. Manuf. Process. 27 (5),
not need of covering fluxes, as it is commonly required after rotary
556–560. https://doi.org/10.1080/10426914.2011.593234.
degassing. Castings produced with HSMC degassed melts exhibit very Campbell, J., 2003. ‘Castings’, 2nd ed. Butterworth-Heinemann, Oxford.
low level of porosity and significantly improved tensile properties, even Chaijaruwanich, A., Dashwood, R.J., Lee, P.D., Nagaumi, H., 2006. Pore evolution in a
after holding the melt for up to 60 min. The HSMC degassing process direct chill cast Al-6 wt.% Mg alloy during hot rolling. Acta Mater. 54 (19),
5185–5194. https://doi.org/10.1016/j.actamat.2006.06.029.
could be easily implemented in industry by replacing the current rotary Dispinar, D., Campbell, J., 2004. Critical assessment of reduced pressure test. Part 1:
impeller technique with minimal impact on the casting process but with porosity phenomena. Int. J. Cast Met. Res. 17 (5), 280–286. https://doi.org/
all the benefits of improved melt quality at reduced cost. 10.1179/136404604225020696.
Dispinar, D., Akhtar, S., Nordmark, A., Di Sabatino, M., Arnberg, L., 2010. Degassing,
hydrogen and porosity phenomena in A356. Mater. Sci. Eng. A 527, 3719–3725.
CRediT authorship contribution statement https://doi.org/10.1016/j.msea.2010.01.088.
Dong, X., Zhang, Y., Amirkhanlou, S., Ji, S., 2018. High performance gravity cast
Al9Si0.45Mg0.4Cu alloy inoculated with AlB2 and TiB2. J. Mater. Process. Technol.
Jaime Lazaro-Nebreda: Methodology, Investigation, Validation, 252, 604–611. https://doi.org/10.1016/j.jmatprotec.2017.10.028.
Formal analysis, Writing - original draft, Visualization. Jayesh B. Patel: Dybalska, A., Eskin, D., Patel, J.B., 2017. Evaluation of shearing time sufficient for
Conceptualization, Investigation, Resources, Writing - review & editing, effective liquid metal processing. JOM 69, 720–724. https://doi.org/10.1007/
s11837-017-2286-x.

11
J. Lazaro-Nebreda et al. Journal of Materials Processing Tech. 294 (2021) 117146

Eskin, D., Alba-Baena, N., Pabel, T., da Silva, M., 2015. Ultrasonic degassing of Tzanakis, I., Lebon, G.S.B., Eskin, D.G., Pericleous, K.A., 2017. Characterizing the
aluminium alloys: basic studies and practical implementation. Mater. Sci. Technol. cavitation development and acoustic spectrum in various liquids. Ultrason.
31 (1), 79–84. https://doi.org/10.1179/1743284714Y.0000000587. Sonochem. 34, 651–662. https://doi.org/10.1016/j.ultsonch.2016.06.034.
Fan, Z., Wang, Y., Xia, M., Arumuganathar, S., 2009. Enhanced heterogeneous nucleation Uludag, M., Cetin, R., Gemi, L., Dispinar, D., 2018. Change in porosity of A356 by
in AZ91D alloy by intensive melt shearing. Acta Mater. 57 (16), 4891–4901. https:// holding time and its effect on the mechanical properties. J. Mater. Eng. Perform. 27,
doi.org/10.1016/j.actamat.2009.06.052. 5141–5151. https://doi.org/10.1007/s11665-018-3534-0.
Fan, Z., Zuo, Y.B., Jiang, B., 2011. A new technology for treating liquid metals with Wan, B., Chen, W., Mao, M., Fu, Z., Zhu, D., 2018. Numerical simulation of a stirring
intensive melt shearing. Mater. Sci. Forum 690, 141–144. https://doi.org/10.4028/ purifying technology for aluminum melt. J. Mater. Process. Technol. 251, 330–342.
www.scientific.net/msf.690.141. https://doi.org/10.1016/j.jmatprotec.2017.09.001.
Gyarmati, G., Fegyverneki, G., Tokár, M., Mende, T., 2020. The effects of rotary Wang, Y., Neff, D., Schwam, D., Zhu, X., Chen, C., 2013. Optimization of permanent mold
degassing treatments on the melt quality of an Al–Si casting alloy. Int. J. Met. mechanical property test bars in A356 alloy using a new mold design. Int. J. Met. 7,
https://doi.org/10.1007/s40962-020-00428-z. 25–38. https://doi.org/10.1007/BF03355556.
Hakansson, A., 2018. Rotor-Stator mixers: from batch to continuous mode of operation-A Warke, V.S., Shankar, S., Makhlouf, M.M., 2005a. Mathematical modelling and computer
review. Processes 6 (4), 32. https://doi.org/10.3390/pr6040032. simulation of molten aluminum cleansing by the rotating impeller degasser: part I.
Hernández-Hernández, M., Cruz-Mendez, W., González-Rivera, C., Ramírez-Argáez, M. Fluid flow. J. Mater. Process. Technol. 168 (1), 112–118. https://doi.org/10.1016/j.
A., 2015. Effect of process variables on kinetics and gas consumption in Rotor- jmatprotec.2004.10.017.
Degassing assisted by physical and mathematical modeling. Mater. Manuf. Process. Warke, V.S., Shankar, S., Makhlouf, M.M., 2005b. Mathematical modelling and computer
30 (2), 216–221. https://doi.org/10.1080/10426914.2014.952303. simulation of molten aluminum cleansing by the rotating impeller degasser: part II.
Hernández-Hernández, M., Camacho-Martínez, J.L., González-Rivera, C., Ramírez- Removal of hydrogen gas and solid particles. J. Mater. Process. Technol. 168 (1),
Argáez, M.A., 2016. Impeller design assisted by physical modeling and pilot plant 119–126. https://doi.org/10.1016/j.jmatprotec.2004.10.016.
trials. J. Mater. Process. Technol. 236, 1–8. https://doi.org/10.1016/j. Yamamoto, T., Kato, K., Komarov, S.V., Ueno, Y., Hayashi, M., Ishiwata, Y., 2018.
jmatprotec.2016.04.031. Investigation of melt stirring in aluminum melting furnace through Water model.
Kang, Y.Y., Lin, Y., Liu, X.D., Sun, C., Yuan, S.S., Zuo, Y.B., Cui, J.Z., 2015. Study on the J. Mater. Process. Technol. 259, 409–415. https://doi.org/10.1016/j.
high shear degassing process with water simulation. Adv. Mat. Res. 1120–1121, jmatprotec.2018.04.025.
1214–1219. https://doi.org/10.4028/www.scientific.net/amr.1120-1121.1214. Zhang, L., Lv, X., Torgerson, A.T., Long, M., 2011. Removal of impurity elements from
Lazaro-Nebreda, J., Patel, J.B., Scamans, G., Fan, Z., 2018. Multi-purpose high shear melt molten aluminum: a review. Miner. Process. Extr. Metall. Rev. 32 (3), 150–228.
conditioning technology for effective melt quality and for recycling of Al-alloy scrap. https://doi.org/10.1080/08827508.2010.483396.
In: Proceedings of the 16th International Aluminium Alloys Conference (ICAA16). Zhang, Y., Patel, J.B., Lazaro-Nebreda, J., Fan, Z., 2018. Improved defect control and
17-21 June, Montreal, Canada, p. 401623. http://www.icaa-conference.net/ mechanical property variation in high-pressure die casting of A380 alloy by high
ICAA16/Papers/Sustainability/401623%20Lazaro-Nebreda_final.pdf. shear melt conditioning. JOM 70, 2726–2730. https://doi.org/10.1007/s11837-
Lordan, E., Lazaro-Nebreda, J., Zhang, Y., Fan, Z., 2019. Effective degassing for reduced 018-3005-y.
variability in high-pressure die casting performance. JOM 71, 824–830. https://doi. Zhang, Y., Lordan, E., Duo, K., Wang, S., Fan, Z., 2020. Influence of porosity
org/10.1007/s11837-018-3186-4. characteristics on the variability in mechanical properties of high pressure die
Lordan, E., Lazaro-Nebreda, J., Zhang, Y., Duo, K., Blake, P., Fan, Z., 2020. On the casting (HPDC) AlSi7MgMn alloys. J. Manuf. Process. 56A, 500–509. https://doi.
relationship between internal porosity and the tensile ductility of aluminium alloy org/10.1016/j.jmapro.2020.04.071.
die-castings. Mater. Sci. Eng. A 778, 139107. https://doi.org/10.1016/j. Zuo, Y., Jiang, B., Enright, P., Scamans, G.M., Fan, F., 2011. Degassing of LM24 Al alloy
msea.2020.139107. by intensive melt shearing. Int. J. Cast Met. Res. 24, 307–313. https://doi.org/
Mostafaei, M., Ghobadi, M., Eisaabadi, B., G, Uludag, M., Tiryakioglu, M., 2016. 10.1179/1743133611Y.0000000002.
Evaluation of the effects of rotary degassing process variables on the quality of A357 Zuo, Y.B., Jiang, B., Zhang, Y., Fan, Z., 2013. Degassing LM25 aluminium alloy by novel
aluminum alloy castings. Metall. Mater. Trans. B 47, 3469–3475. https://doi.org/ degassing technology with intensive melt shearing. Int. J. Cast Met. Res. 26 (1),
10.1007/s11663-016-0786-7. 16–21. https://doi.org/10.1179/1743133612Y.0000000019.
Polmear, I.J., 1995. Light Alloys: Metallurgy of the Light Metals, 3rd ed. Virginia Zuo, Y.B., Kang, Y.Y., Lin, Y., Liu, X., Sun, C., Yuan, S., Cui, J., 2015. A new high shear
University: Wiley. degassing technology and mechanism for 7032 alloy. China Foundry 12, 293–298.
Scargiali, F., Busciglio, A., Grisafi, F., Brucato, A., 2014. Bubble formation at variously Zuo, Y.B., Kang, Y.Y., Lin, Y., Zhu, Q.F., Li, L., Li, Z.Z., Cui, J.Z., 2016a. Dispersion
inclined nozzles. Chem. Eng. Technol. 37 (9), 1507–1514. https://doi.org/10.1002/ behaviour of Ar bubbles under intensive shearing and its effect on degassing effect of
ceat.201300511. 7075 alloy. Chinese J. Nonferrous Metals 26 (3), 486–493. https://caod.oriprobe.co
Tan, E., Tarakcilar, A., Dispinar, D., Colak, M., Kayikci, R., 2011. Reproducibility of m/articles/48132534/Dispersion_behaviour_of_Ar_bubbles_under_intensive.htm.
reduced pressure test results in testing of liquid aluminum gas levels. In: Proceedings Zuo, Y.B., Lin, Y., Kang, Y.Y., Cui, J.Z., 2016b. effects of rotor rotation speed and gas flow
of the 6th International Advanced Technologies Symposium (IATS11). 16-18 May, rate on the degassing efficiency of 2524 aluminum alloy in rotary degassing process.
Elazig, Turkey, pp. 321–324. J. of Northeastern Univ. Nat. Sci. 37 (5), 653–657. http://xuebao.neu.edu.cn
Tiryakioglu, M., 2020. The effect of hydrogen on pore formation in aluminum alloy /natural/EN/10.12068/j.issn.1005-3026.2016.05.010.
castings: myth versus reality. Metals 10 (3), 368. https://doi.org/10.3390/
met10030368.

12