Wettability Effect on Bubble Formation at Nozzles in Liquid Aluminum

Materials Transactions, Vol. 44, No. 11 (23) pp. 2298 to 232 #23 The Japan Institute of Metals Wettability Effect on Bubble Formation at Nozzles in Liquid Aluminum Svyatoslav V. Gnyloskurenko 1; * and Takashi Nakamura 2 1 Institute of Industrial Science, The University of Tokyo, Tokyo 153-855, Japan 2 Institute of Multidisciplinary Research for Advanced Materials, Tohoku University, Sendai 98-8577, Japan An experimental study was undertaken to determine how wettability affect the volume and formation of gas bubble at nozzles in liquid aluminum. X-ray fluoroscope was used to carry out in-situ observation in the melt. The nozzles were made of steel, silica and alumina to establish different wettability by liquid aluminum. The frequency of bubble formation and the bubbles volume were investigated for low gas flow rate from.43 to 12 cm 3 /s and for various diameters of the nozzle. It was shown that bubble volume increased with wettability worsening both for aqueous and metallic systems. A further insight into the mechanism of the bubble formation was obtained by comparison of the bubble behavior at the tip of the injection devices in liquid aluminum and water. (Received July 4, 23; Accepted September 16, 23) Keywords: wettability, contact angle, bubble formation, bubble volume, nozzle, gas injection, foamed metal 1. Introduction The introduction of gaseous phase into liquid metal bath is specific feature of many metallurgical processes, refining operations and melt foaming techniques. The size of bubbles plays a key role in controlling both the chemical kinetics and mass transfer during melt treatment or in determining the structure of cellular metals. Although bubble formation in aqueous and metallic systems have been experimentally studied for years, 1 6) this process is considered to be extremely complicated phenomenon due to a variety of systems and operating variables such as gas flow rate, nozzle diameter, volume of the gas chamber, liquid and gas properties. Despite the different approaches and the discrepancies in the results it has been found that the bubble formation at the gas-liquid-solid interfaces (tip of nozzle etc.) at low gas flow rate, Q is substantially controlled by surface phenomena. The bubble volume at the moment of release from the nozzle, V b is determined by a balance between the buoyancy force on the one hand and on the other surface tension forces constraining the bubble to the inner diameter of the nozzle, D i. 3,4,7) Hence V b ¼ D i =ð g Þg ð1þ where and g the densities of liquid and gas, correspondingly, the surface tension of the liquid-gas interface, g the acceleration due to gravity. These predicted bubble volumes apply to the systems in which the liquid wets the nozzle material. However, for nonwetting systems bubbles generally tend to form at the outer diameter of the nozzle, D o so that the bubble volume becomes V b ¼ D o =ð g Þg ð2þ The bubble volume remains constant independently of moderate changes in the gas-flow rate. So, the bubbling frequency is proportional to the gas flow rate in this range, *Corresponding author, E-mail: slava@iis.u-tokyo.ac.jp named the region of constant volume or static bubbles. As the gas flow rate further increases the bubble frequency levels off at a constant value and thereafter the bubble volume rises proportionally to the gas flow rate until a continuous jet occurs. To refine eqs. (1) and (2) a capacitance number, N c relating the acoustical properties of the gas-nozzle system to the chamber volume, V c was included by many researchers. 3,4) Gas chamber volume is defined as the volume between the last large pressure drop and the nozzle tip. Irons and Guthrie 4) have derived N c ¼ 4gV c =D 2 i P, where P the static pressure over the nozzle tip. Depending on N c bubble volume V at low flow rates can be expressed as V ¼ V b (from eq. (1)) at < N c < 1; V ¼ N c V b at 1 < N c < 9 and V ¼ 9:14V b at N c > 9. Sano and Mori 3) have developed a capacitance number N c ¼ 4gV c sin =D i D o P, where contact angle and sin is assumed to be unity in the absence of any measured contact angles. Although, the effect of nozzle wettability by liquid, i.e. contact angle, between solid surface, on which bubble is formed and surrounding fluid medium, to the bubble volume and formation mechanism was well studied for aqueous systems 8 11) a direct extrapolation of the results to liquid metals is often misleading. A few studies in liquid metal showed that wettability of nozzles plays an important role in bubble growth. 3,4,12) Authors 3,12) observed that the outer diameter of non-wetted nozzles controls the bubble size in liquid mercury and silver. Irons and Guthrie 4) confirmed this result for liquid iron. Ogino and Nishiwaki 13) experimentally proved bubble decrease with wettability enhancement for the CO bubble - liquid slag - solid SiC system. Studies on interfacial phenomena in slag-metal-gas systems 14) and in metal foaming 15) also revealed the same tendency. However, bubble growth mechanism at varying wettabilty in high temperature melts is still unclear. The present study aims to investigate in situ the effect of wettability during the injection of argon into molten aluminum on bubble formation mechanism and volume at detachment at the nozzle tip of different material and wetting. Wide range of contact angles was realized in liquid

Wettability Effect on Bubble Formation at Nozzles in Liquid Aluminum 2299 Amplifier Thermocouple Flow meter Ar cylinder (b) Gas outlet X-ray controller Temperature monitoring CCD, Intensifier X-ray tube Injection pipe Inner diameter,d i Monitor Outer diameter,d o Generator Video tape recorder Furnace 2mm Fig. 1 Schematic view of the X-ray apparatus for bubble formation study and the sketch of the nozzle (b). aluminum using nozzles of steel, silica and alumina to establish different wettability by the melt. X-ray fluoroscope was used to observe gas bubble evolution in the melt. The frequency of bubble formation and the volume of the bubbles, V b were investigated for low gas flow rate, Q ¼ :43 to 12 cm 3 /s and for various diameters of the nozzle, D i ¼ :1{:4 cm and D o ¼ :2{1: cm. It was shown that bubble volume increased with contact angle. A further insight into the mechanism of the bubble formation was obtained by comparison of the present results in liquid aluminum and our previous study in water. 11) 2. Experimental Apparatus and Procedure The X-ray system (Fig. 1) was used to observe and record bubble growing at the nozzle in aluminum melt. An X- ray beam from a 15 kv source (Hitachi Medical Corporation) was passed through a furnace and crucible filled with the melt and formed an image on fluorescent screen placed immediately behind the furnace. The image was transmitted via an intensifier to a CCD camera (XTV-S-42, Hitachi Medical Corporation), displayed on a monitor and recorded either continuously onto videotape or periodically onto PC disk for image enhancement. The nozzles facing upward were used to inject gas into the melt (Fig. 1(b)). Injection devises of.1 to.4 cm in inner diameter (I.D.) and of.2 to 1 cm in outer diameter (O.D.) were attached to steel pipe of.8 m in length by Sumiceram alumina paste and then to an acrylic pipe of 5 m connected to gas cylinder via flow rate controller and pressure gauge. The gas chamber volume was equaled 73 cm 3 that is much large then the volume of each bubble released. So, the experiments were carried out in constant pressure conditions. 4) The nozzles were made of steel, silica and alumina to establish different wettability by liquid aluminum. In the present work the term good or poor wettability was used with reference to rough values of the contact angle, because precise high-temperature measurement of interfacial phenomena usually meets significant difficulties. Since iron is perfectly wetted by aluminum ( 9 ), 16) 33, 17) low carbon steel was chosen as a nozzle material, performing good wettability. Nonreactive liquid metal-oxide systems exhibit poor wettability and the contact angle typically is larger than 12, 18) which also holds for Al 2 O 3 in liquid aluminum. 19) So, it was considered here that an alumina nozzle leads to poor wettability. Authors 2) reported the contact angle for SiO 2 in liquid Al to decrease from 15 to 9 with time and temperature increase, whereas Kaptay 21) indicated that ¼ 5{6. So, wettability of silica nozzle by aluminum is considered here to be intermediate, i.e. better than that of alumina and poorer than that of steel. A 1 g of pure aluminum (99.9 mass%) grains placed in an alumina crucible with a rectangular cross section of 15 mm 36 mm were melted in a resistance furnace under ambient atmosphere. When the prescribed temperature of the melt (69 C) was reached the injection device was preheated in a furnace and introduced vertically into the melt to a depth of 5 cm while argon gas blowing. The flow rate was increased step by step from.43 to 12 cm 3 /s and X-ray observation was carried out for 3 s at every constant flow rate. When bubbling frequency significantly increased and metal splashing became excessive experiment was finished. The frequency of bubble formation, f was visually evaluated by counting the number of bubbles generated at the nozzle exit on video tape. The bubble volume V b,cm 3 was then determined by using equation V b ¼ Q=f, where Q is the gas flow rate, cm 3 /s. 3. Results and Discussion The results of bubble formation study in liquid aluminum are presented in Figs. 2 5.

23 S. V. Gnyloskurenko and T. Nakamura Bubble Volume, Vb /cm 3 1.9.8.7.6.5.4.3.2.1 2/1 mm 2.5/1.5 mm 3/2 mm 4/3 mm 5/4 mm 1 2 3 4 5 6 Gas Flow Rate, Q/ cm 3. s (c) (e) Bubble Volume, Vb / cm 3 1.1 1.9.8.7.6.5.4.3.2.1 (b) Steel Nozzle, 2/1 mm 1/1 mm Silica Nozzle, 2/1mm 1/1mm Alumina Nozzle, 2/1mm 1/1mm 1 2 3 4 5 6 Gas Flow Rate, Q/ cm 3.s Fig. 2 Bubble volume dependence on gas flow rate for steel nozzles of different diameter and (b) nozzles of different materials, outer diameter and equal inner diameter (D i ¼ 1 mm). (b) 1mm (d) (f) Fig. 3 X-ray images of bubble formed at the nozzle of 4 mm I.D. and 1 mm O.D. made of, (b) steel, (c), (d) silica and (e), (f) alumina., (c) and (e) - bubble formation at the tip after emergence from the hole; (b), (d) and (f) - bubble growth and floating start. (b) (c) (d) (e) (f) 1 mm Water - Air system (g) (h) (i) (j) (k) (l) 5 mm Liq. Al - Argon system Bubble formation stage Wetting conditions, systems Growth Good wettability ( θ <9 ) Necking and Detachment Water Acrylic Plastic Air, (c) Steel nozzle Liquid Aluminum Argon, (g) (i) Growth Poor wettability ( θ >9 ) Necking and Detachment Water- Acrylic Plastic coated by paraffin Air, (d) (f) Alumina nozzle Liquid Aluminum Argon, (j) (l) Fig. 4 Scheme of bubble formation in aqueous 11) and metallic systems of different wetting. Figure 2 shows that bubble volume increases from.16 to.8 cm 3 with nozzle diameter that is also reflected by eqs. (1) and (2). Bubble volume depends weakly on gas flow rate proving that the bubbles were formed in the constant volume region. Irons and Guthrie 4) have also shown similar bubbling pattern in liquid iron at gas flow rate less than 12 cm 3 /s. In

Wettability Effect on Bubble Formation at Nozzles in Liquid Aluminum 231 Bubble Volume, Vb, cm 3.6.5.4.3.2.1 Water-Air system, 11) Q=.33 cm 3. s 1, 24) Data of Mukai et al., Q=1 cm 3. s, D i =.8 mm Liq. Al-Argon (present data), Q=.648 cm 3. s Liq. Al-Argon, Q=.864 cm 3. s 4 5 6 7 8 9 1 11 12 13 14 15 Contact Angle, Θ Fig. 5 Dependence of bubble volume, V b on contact angle, in Water- Air 1,11,24) and Liquid Aluminum - Argon systems. Q - gas flow rate, inner diameter of the injection hole, D i ¼ 1 mm. this region the bubble volume is almost independent of flow rate and surface forces have the predominant effect on the bubble formation. It is apparent from Fig. 2 that bubble volume issued from larger nozzles decreases at the beginning of gas injection (descending part of the curves). It may be attributed to the fact that bubble growth and flotation within initially motionless melt resulted in flow of liquid in the crucible that may facilitate bubble formation and detachment. As the gas flow rate was further increased up to certain value, the melt flow was stabilized and bubbles of equal volume were formed (plateau part of the curves, Fig. 2). It was observed that the rising bubbles did not touch the crucible wall. When gas flow rate exceeded the range of 2.5 6.5 cm 3 /s in dependence on diameter and nozzle material, bubble frequency was significantly increased (14 s 1 ) and no vertical displacement between successive bubbles was possible to observe. Finally, gas issued continuously from the nozzle with no separated bubbles and jet mode was considered to occur. Experimental bubble volumes obtained for steel nozzles of different diameter were compared with those ones calculated by eqs. (1), (2) (Table 1). Gas chamber effect ( capacitance numbers, N c 4) and N c 3) ) was also considered. Experimental values represented volume of the bubbles issued at the stable bubbling mode (plateau part of the curves, Fig. 2). Apparently, data obtained by eq. (1) are best fitted to the experimental bubble volume. So, it is reasonable to conclude that bubbles were formed at the inner diameter of the nozzle. Calculation with capacitance numbers fitted data to the experimental results only at large nozzle diameter. To reveal bubble behavior on the nozzle tip of different wettability the experiments with nozzles of steel, silica and alumina, equal inner diameter (D i ¼ 1 mm) and different outer ones (D o ¼ 2 and 1 mm) were carried out. Figure 2(b) shows that alumina nozzle produced much large bubbles that steel and silica nozzles did confirming the role of wettability. Injection devices of 2 mm in O.D. formed much smaller bubbles than nozzles of 1 mm did meaning that bubbles first spreaded toward the outer circumference for all non-wetted nozzles and then detached. Tip of the steel nozzles supposed to be wetted by molten aluminum was probably partly oxidized while preheating in the furnace, which caused wettability worsening. Therefore, bubble was most likely spreaded over oxidized steel tip and its volume also increased with outer diameter of the nozzle (Fig. 2(b)). The role of steel oxidation on bubble formation was qualitatively estimated by using steel nozzle preheated for 15 min in a furnace and then introduced into the melt and blew gas. For comparison the steel nozzle without preheating (non-oxidized) was lowered into the melt immediately while gas blowing. X-ray observation showed that bubble grew while expanding towards the outer edge of the oxidized nozzle, and conversely formed around the hole over nonoxidized nozzle tip. Apparently, the last nozzle issued much smaller bubble than the former did. X-ray apparatus allowed observing in situ bubble spread over the nozzle tip of different material in liquid melt (Fig. 3). It is clear that the bubble emerged from the hole was spreaded to the outer diameter of the non-wetted alumina nozzle (Fig. 3(e)). In contrast to this, the bubble was spreaded to a limited extend smaller (6 mm) than the outer diameter (1 mm) over the nozzle tip of steel (Fig. 3). So, bubble produced by steel and silica nozzles were smaller than that formed on alumina tip (poorer wettability) that was confirmed both graphically (Fig. 2(b)) and qualitatively (Fig. 3). The larger adherence area was the more extended time was needed to establish conditions for bubble detachment from a nozzle (force balance). During this period bubble was supplied with an additional gas with increase in the bubble volume at detachment. Moreover the contact angles also controlled adherence force 8) and they were supposed to increase (wettability worsening) from steel to alumina material in aluminum also contributing to bubble volume increase. In situ observation made it possible to compare a bubble Table 1 Experimental and calculated bubble volume obtained with steel nozzles. Nozzle Bubble volume, cm 3 diameter Present Calculated D i, 1 1 cm study V b = V b = N c V b (eq. (1)) V b (eq. (1)) D i =ð g Þg D o =ð g Þg or or Eq. (1) Eq. (2) 9:14 V b N c V b V b at N c V b at N c 1.166.124.247 1.133.719 2.27.247.371 1.39.469 3.31.371.49.89.371 4.49.49.62.637.49

232 S. V. Gnyloskurenko and T. Nakamura formation mechanism at the tip of the injection devices in liquid melt and in water. Figure 4 depicts the scheme of a bubble evolution in metallic and aqueous systems of different wetting. The images of the air bubble were taken from the aqueous experiments at low gas flow rate and injection hole diameter of 1 mm. 11) Wettability at the air-water-acrylic plastic interface was considered as good wettability for uncoated plastic, ¼ 68 in Figs. 4 (c)) and poor wettability for plastic coated by paraffin, ¼ 11 in Figs. 4(d) (f)). Good ( <9 ) and poor ( >9 ) wettability in aluminum were represented by steel (Figs. 4(g) (i)) and alumina (Figs. 4(j) (l)) nozzles, respectively. Similar stages of bubble formation at low flow rate were revealed in water and liquid aluminum, named growth and necking while detaching. After arising from the hole gas spreaded to a limited extend over the wetted tip (Figs. 4 and (g)), whereas it covered larger adherence area over nonwetting surface (Figs. 4(d) and (j)). This stage can be corresponded to the under critical stage of the bubble evolution in water. 11) The bubble further grew only upward (Figs. 4(h) and (k)) as happened during critical growth stage in water (Figs. 4(b) and (e)) 11) when bubble shape deformed and became elongated. Similar bubble behavior and shape was also observed in aqueous system 8,9,22,23) and mercury. 14) During the last stage the neck formation began and bubble size formed under poor wetting conditions (Figs. 4(f) and (l)) was found to be much larger than that under good wettability (Figs. 4(c) and (i)). So, at very low gas flow rate bubble formed over the solid surface in liquid aluminum and water in similar way under predominant effect of wettability. Figure 5 shows the dependence of the bubble volume on contact angle (wettability of injection devices, D i ¼ 1 mm) in liquid aluminum and water. 1,11,24) Experimental bubble volume in metallic system are presented for the lowest gas flow rates (Q ¼ :648 and.864 cm 3 /s) used in the present study. It was determined that bubble volume increased with contact angle both for aqueous and metallic systems. Apparently, bubbles formed in water 11) are one order of magnitude smaller (.3.6 cm 3 ) than that in liquid aluminum (.17.49 cm 3 ). This can be mainly attributed to the significant difference in surface forces taking place in aqueous and metallic systems and gas flow rates applied for bubble formation study. Data of Mukai et al. for aqueous system 1,24) (Fig. 5) confirms that increase in gas flow rate up to 1 cm 3 /s resulted in bubble volume increase and fitted the data to that of the present metallic system. 4. Conclusions In situ bubble formation at the nozzles in liquid aluminum was studied by using X-ray fluoroscopic technique. Experiments carried out under low gas flow rate using nozzles of different material showed the influence of wettability on bubble volume and formation mechanism. It was determined that a bubble produced by steel and silica nozzles was smaller than that formed from an alumina tip of poorer wettability. It was shown that bubble volume increased with contact angle both for aqueous and metallic systems. A further insight into the mechanism of bubble formation was obtained by comparison of bubble behavior at the tip of the injection devices in the liquid melt and water. The examination revealed similar stages of bubble evolution, named growth and necking while detaching. Bubbling mode was indicated to turn into jet mode when gas flow rate exceeded the range of 2.5 6.5 cm 3 /s depending on diameter and nozzle material. The present study showed that injection devices wetted by liquid melt are reasonable to use, particularly in metal foaming techniques, for producing small bubbles in the melt. REFERENCES 1) L. Davidson and E. H. Amick: AIChE J. 2 (1956) 337 342. 2) J. F. Davidson, A. M. I. Mech and B. O. G. Shuler: Trans. Inst. Chem. Eng. 38 (196) 335 342. 3) M. Sano and K. Mori: Trans., JIM 17 (1976) 344 352. 4) G. A. Irons and I. L. Guthrie: Metall. Trans. 9B (1978) 11 11. 5) M. Iguchi, M. Kaji and Z.-I. Morita: Metall. Mater. Trans. B 29B (1998) 129 1218. 6) S. V. Gnyloskurenko, T. Nakamura and Y. Waseda: Proc. 2nd Int. Conf. on Processing Materials for Properties, ed. by Br. Mishra and Ch. Yamauchi, (TMS, San Francisco, 2) pp. 49 52. 7) R. J. Benzing and J. E. Myers: Ind. Eng Chem. Engineering, Design and Equipment 47 (1955) 287 29. 8) I. W. Wark: Journ. Physic. Chem. 37 (1933) 623 644. 9) V. W. Fritz: Physik. Zeitschr. 36 (1935) 379 384. 1) K. Mukai, H. Nozaki and T. Arikawa: CAMP ISIJ 3 (199) 137. 11) S. V. Gnyloskurenko, A. V. Byakova, O. I. Raychenko and T. Nakamura: Colloids Surf., A 218 (23) 73 87. 12) K. Okumura, R. Harris and M. Sano: Can. Metall. Q. 37 (1998) 49 56. 13) K. Ogino and T. Nishiwaki: Tetsu to Hagane 65 (1979) 1985. 14) H. Terashima, T. Nakamura, K. Mukai and D. Izu: J. Japan Inst. Metals 56 (1992) 422 429. 15) T. Nakamura, S. V. Gnyloskurenko, K. Sakamoto, A. V. Byakova and R. Ishikawa: Mater. Trans. 43 (22) 1191 1196. 16) Yu. V. Naydich: Contact phenomena in metallic melts, (Kiev, Naukova Dumka, 1972) p. 196 (in Russian). 17) V. N. Eremenko, T. S. Ivanova and N. D. Lesnik: Adhesion of melts and welding of materials 6 (198) 51 54 (in Russian). 18) B. M. Gallois: JOM June (1997) 48 51. 19) V. Laurent, D. Chatain, C. Chatillon and N. Eustathopoulos: Acta Metall. 36 (1988) 1797 183. 2) V. Laurent, D. Chatain and N. Eustathopoulos: Mater. Sci. Eng. A135 (1991) 89 94. 21) G. Kaptay: Int. Conf. on Metal Foams and Porous Metal Structure, ed. by J. Banhart, M. F. Ashby and N. A. Fleck (MIT-Verlag, Bremen, 1999) pp. 141 145. 22) B. Kabanov and A. Frumkin: Z. Physik. Chem. (A) Band 165 (1933) 433 452. 23) Y. Mizuno and M. Iguchi: ISIJ Int. 41 (21) S56-S6. 24) K. Mukai: ISIJ Int. 32 (1992) 19 25.